University of Iowa
Iowa Research Online
Theses and Dissertations
2008
How do teams learn? shared mental models and
transactive memory systems as determinants of
team learning and effectiveness
Amit Kumar Nandkeolyar
University of Iowa
Copyright 2008 Amit Kumar Nandkeolyar
This dissertation is available at Iowa Research Online: http://ir.uiowa.edu/etd/2
Recommended Citation
Nandkeolyar, Amit Kumar. "How do teams learn? shared mental models and transactive memory systems as determinants of team
learning and effectiveness." PhD (Doctor of Philosophy) thesis, University of Iowa, 2008.
http://ir.uiowa.edu/etd/2.
Follow this and additional works at: http://ir.uiowa.edu/etd
Part of the Business Administration, Management, and Operations Commons
HOW DO TEAMS LEARN? SHARED MENTAL MODELS AND TRANSACTIVE
MEMORY SYSTEMS AS DETERMINANTS OF TEAM LEARNING AND
EFFECTIVENESS
by
Amit Kumar Nandkeolyar
An Abstract
Of a thesis submitted in partial fulfillment
of the requirements for the Doctor of
Philosophy degree in Business Administration
in the Graduate College of
The University of Iowa
August 2008
Thesis Supervisor: Professor Greg L. Stewart
1
ABSTRACT
Shared mental models (SMM) and Transactive memory systems (TMS) have been
advocated as the main team learning mechanisms. Despite multiple appeals for
collaboration, research in both these fields has progressed in parallel and little effort has
been made to integrate these theories. The purpose of this study was to test the
relationship between SMM and TMS in a field setting and examine their influence on
various team effectiveness outcomes such as team performance, team learning, team
creativity, team members’ satisfaction and team viability.
Contextual factors relevant to an organizational setting were tested and these
included team size, tenure, country of origin, team reward and organizational support.
Based on responses from 41 teams from 7 industries across two countries (US and India),
results indicate that team size, country of origin and team tenure impact team
performance and team learning. In addition, team reward and organizational support
predicted team viability and satisfaction.
Results indicated that TMS components (specialization, coordination and
credibility) were better predictors of team outcomes than the omnibus TMS construct. In
particular, TMS credibility predicted team performance and creativity while TMS
coordination predicted team viability and satisfaction. SMM was measured in two
different ways: an average deviation index and a 6-item scale. Both methods resulted in a
conceptually similar interpretation although average deviation indices provided slightly
better results in predicting effectiveness outcomes.
TMS components moderated the relationship between SMM and team outcomes.
Team performance was lowest when both SMM and TMS were low. However, contrary
to expectations, high levels of SMM did not always result in effective team outcomes
2
(performance, learning and creativity) especially when teams exhibited high TMS
specialization and credibility. An interaction pattern was observed under conditions of
low levels of SMM such that high TMS resulted in higher levels of team outcomes. The
theoretical and practical implications of these results are discussed.
Abstract Approved: ____________________________________
Thesis Supervisor
____________________________________
Title and Department
____________________________________
Date
HOW DO TEAMS LEARN? SHARED MENTAL MODELS AND TRANSACTIVE
MEMORY SYSTEMS AS DETERMINANTS OF TEAM LEARNING AND
EFFECTIVENESS
by
Amit Kumar Nandkeolyar
A thesis submitted in partial fulfillment
of the requirements for the Doctor of
Philosophy degree in Business Administration
in the Graduate College of
The University of Iowa
August 2008
Thesis Supervisor: Professor Greg L. Stewart
Copyright by
AMIT KUMAR NANDKEOLYAR
2008
All Rights Reserved
Graduate College
The University of Iowa
Iowa City, Iowa
CERTIFICATE OF APPROVAL
_______________________
PH.D. THESIS
_______________
This is to certify that the Ph.D. thesis of
Amit Kumar Nandkeolyar
has been approved by the Examining Committee
for the thesis requirement for the Doctor of Philosophy
degree in Business Administration at the August 2008 graduation.
Thesis Committee: ___________________________________
Greg L. Stewart, Thesis Supervisor
___________________________________
Terry L. Boles
___________________________________
Kenneth G. Brown
___________________________________
Roy Suddaby
___________________________________
Jennifer Glanville
ACKNOWLEDGMENTS
I am immensely grateful for the encouragement and support of many people who
have helped me successfully complete this dissertation. Their contribution is
immeasurable and I would like to mention my heartfelt gratitude for their efforts and
kindness.
I would like to express my deep gratitude to my adviser, Dr. Greg Stewart, whose
motivation, patience and tireless energy helped sustain me through the highs and the lows
of this project. I am deeply indebted to Greg for encouraging and mentoring me
throughout my PhD and this project.
Next, I would like to thank the other members of my dissertation committee: Dr.
Terry Boles, Dr. Ken Brown, Dr. Roy Suddaby and Dr. Jennifer Glanville. All of them
provided me with important insights, unfailing support and valuable advice.
Finally, I would like to thank my family and friends who have been a great source
of inspiration and support. In fact, I would like to dedicate this dissertation to my wife,
Aparna, without whom it would have been impossible to accomplish this goal.
ii
ABSTRACT
Shared mental models (SMM) and Transactive memory systems (TMS) have been
advocated as the main team learning mechanisms. Despite multiple appeals for
collaboration, research in both these fields has progressed in parallel and little effort has
been made to integrate these theories. The purpose of this study was to test the
relationship between SMM and TMS in a field setting and examine their influence on
various team effectiveness outcomes such as team performance, team learning, team
creativity, team members’ satisfaction and team viability.
Contextual factors relevant to an organizational setting were tested and these
included team size, tenure, country of origin, team reward and organizational support.
Based on responses from 41 teams from 7 industries across two countries (US and India),
results indicate that team size, country of origin and team tenure impact team
performance and team learning. In addition, team reward and organizational support
predicted team viability and satisfaction.
Results indicated that TMS components (specialization, coordination and
credibility) were better predictors of team outcomes than the omnibus TMS construct. In
particular, TMS credibility predicted team performance and creativity while TMS
coordination predicted team viability and satisfaction. SMM was measured in two
different ways: an average deviation index and a 6-item scale. Both methods resulted in a
conceptually similar interpretation although average deviation indices provided slightly
better results in predicting effectiveness outcomes.
TMS components moderated the relationship between SMM and team outcomes.
Team performance was lowest when both SMM and TMS were low. However, contrary
to expectations, high levels of SMM did not always result in effective team outcomes
iii
(performance, learning and creativity) especially when teams exhibited high TMS
specialization and credibility. An interaction pattern was observed under conditions of
low levels of SMM such that high TMS resulted in higher levels of team outcomes. The
theoretical and practical implications of these results are discussed.
iv
TABLE OF CONTENTS
LIST OF TABLES............................................................................................................. vi
LIST OF FIGURES ......................................................................................................... viii
CHAPTER 1: INTRODUCTION ........................................................................................1
CHAPTER 2: LITERATURE REVIEW .............................................................................7
Theoretical framework......................................................................................7
Shared Mental Models ....................................................................................10
Transactive Memory Systems.........................................................................16
Relationship between TMM and TMS ...........................................................24
Context............................................................................................................32
CHAPTER 3: METHOD ...................................................................................................43
Procedure ........................................................................................................43
Sample ............................................................................................................44
Aggregation Issues..........................................................................................45
Study Variables and Measures........................................................................47
CHAPTER 4: RESULTS...................................................................................................64
Method 1: Team Mental Model (Average Deviation)....................................65
Method 2: Team Mental Model (Alternate Scale)..........................................74
CHAPTER 5: DISCUSSION...........................................................................................130
Theoretical Implications ...............................................................................132
Managerial Implications ...............................................................................144
Limitations....................................................................................................146
Future Research ............................................................................................149
Conclusion ....................................................................................................151
REFERENCES ................................................................................................................152
APPENDIX A: SURVEY................................................................................................162
To be completed by the team members ........................................................162
To be completed by the team managers .......................................................167
APPENDIX B: INFORMED CONSENT DOCUMENT ................................................170
APPENDIX C: GLOSSARY...........................................................................................173
v
LIST OF TABLES
Table 1.
Sample details ..................................................................................................61
Table 2.
Intraclass Coefficients (ICCs) and within-group agreement (Rwg)
indices ..............................................................................................................62
Table 3.
Confirmatory factor analyses of the scales ......................................................63
Table 4.
Means, Standard Deviations and Inter-correlations among study
variables ...........................................................................................................82
Table 5.
Results of Hierarchical Regression Analyses testing Hypothesis 1 with
team tenure, team size and country of origin as control variables:
TMM Average Deviation.................................................................................85
Table 6.
Results of Hierarchical Regression Analyses testing Hypothesis 1 with
task interdependence, team reward and organizational support as control
variables: TMM Average Deviation ................................................................87
Table 7.
Results of Hierarchical Regression Analyses examining impact of TMS
on team outcomes: TMM Average Deviation .................................................89
Table 8.
Results of Hierarchical Regression Analyses testing Hypothesis 2a and
3: TMM Average Deviation.............................................................................91
Table 9.
Results of Hierarchical Regression Analyses testing Hypothesis 2b and
3: TMM Average Deviation.............................................................................93
Table 10. Results of Hierarchical Regression Analyses testing Hypothesis 2c and
3: TMM Average Deviation.............................................................................95
Table 11. Post-hoc Hierarchical Regression Analyses: TMS specialization and
team outcomes .................................................................................................97
Table 12. Post-hoc Hierarchical Regression Analyses: TMS coordination and
team outcomes .................................................................................................99
Table 13. Post-hoc Hierarchical Regression Analyses: TMS credibility and team
outcomes ........................................................................................................101
Table 14. Results of Hierarchical Regression Analyses testing Hypothesis 1 with
team tenure, team size and country of origin as control variables: TMM
Alternate scale................................................................................................103
Table 15. Results of Hierarchical Regression Analyses testing Hypothesis 1 with
task interdependence, team reward and organizational support as
controls: TMM Alternate scale ......................................................................105
Table 16. Results of Hierarchical Regression Analyses examining impact of TMS
on team outcomes: TMM Alternate Scale .....................................................107
vi
Table 17. Results of Hierarchical Regression Analyses testing Hypothesis 2a and
3: TMM Alternate Scale ................................................................................109
Table 18. Results of Hierarchical Regression Analyses testing Hypothesis 2b and
3: TMM Alternate Scale ................................................................................111
Table 19. Results of Hierarchical Regression Analyses testing Hypothesis 2c and
3: TMM Alternate Scale ................................................................................113
Table 20. Summary of key findings...............................................................................129
Table 21. Conceptual overlap across different literatures on team learning..................133
vii
LIST OF FIGURES
Figure 1. Team-level conceptual model describing relationships between
processes, context and dependent variables.....................................................41
Figure 2. Hypothesized relationship between TMM, TMS and team performance ........42
Figure 3. Interaction between TMM and TMS on team creativity ................................115
Figure 4. Curvilinear impact of TMS specialization on team creativity........................116
Figure 5. Curvilinear impact of TMS coordination on team viability ...........................117
Figure 6. Curvilinear impact of TMS coordination on team satisfaction ......................118
Figure 7. Curvilinear impact of TMS credibility on team performance ........................119
Figure 8. Interaction between TMM and TMS specialization on team creativity.........120
Figure 9. Interaction between TMM and TMS coordination on team performance......121
Figure 10. Interaction between TMM and TMS credibility on team performance .........122
Figure 11. Interaction between TMM and TMS credibility on team learning.................123
Figure 12. Interaction between TMM and TMS credibility on team creativity...............124
Figure 13. Interaction between TMM and TMS on team creativity ................................125
Figure 14. Interaction between TMM and TMS specialization on team creativity .........126
Figure 15. Interaction between TMM and TMS coordination on team viability.............127
Figure 16. Interaction between TMM and TMS credibility on team creativity...............128
viii
1
CHAPTER 1: INTRODUCTION
Knowledge is of two kinds. We know a subject ourselves, or we know where
we can find information upon it.
-
Samuel Johnson (In life of Samuel Johnson, 1791 by James
Boswell)
Knowledge is of paramount importance nowadays since organizations are
increasingly becoming service oriented and dependent on knowledge workers (Blackler,
1995; Davenport & Prusak, 1998). Organizations try to capture the benefits of collective
knowledge assimilation by the formation of teams and task-forces (Edmondson, Bohmer,
& Pisano, 2001; Katzenbach & Smith, 2003). Therefore, one of the primary challenges
facing managers is to encourage knowledge sharing within a team so that team members
can effectively combine their unique knowledge (Nonaka, 1994). Knowledge sharing
takes place mostly through interaction with others and we often rely on others to augment
our knowledge (Simon, 1997). A combination of individual and others’ knowledge is an
ideal means to obtain information, solve problems, and be creative in the workplace.
There has been a lot of research in the past decade on how groups engage in
knowledge creation and sharing. For example, organizational researchers have
demonstrated the importance of learning in groups and communities and its importance to
team outcomes (Argote, Gruenfeld, & Naquin, 2001). Similarly, social psychologists
have proposed that groups develop a shared understanding through a process of
interaction, and the resulting knowledge is shared and distributed amongst group
members (Thompson & Fine, 1999). Strategic Management researchers have proposed a
knowledge-processing view of the firm that emphasizes the importance of social
interaction as the process through which knowledge is created and transferred in
organizations (Kogut & Zander, 1992; Nonaka, 1994). It can benefit all fields to integrate
2
findings on group learning from these disciplines to better understand how group learning
occurs.
I define team learning as the process by which team members seek to acquire,
share, refine, or combine task-relevant knowledge through interaction with one another
(Argote, Gruenfeld, & Naquin, 2001). Such a definition will include asking questions,
challenging assumptions, seeking multiple perspectives and reflecting on past actions
(Van der Vegt & Bunderson, 2005). The team learning literature has focused on how
successful groups reach better decisions by assimilating the knowledge residing in
individual members.
Team learning represents a process term under the Input-Process-Outcome (or IP-O) framework (Hackman, 1987). These processes are mediating mechanisms linking
input variables such as member, team, and organizational characteristics with output
measures like team performance, team members’ satisfaction and team viability. In
addition to input variables, the team learning process is likely to be influenced by
contextual factors (input variables) external to the team. Situational factors have been
underappreciated in organizational research and need to be better documented to improve
our understanding of the phenomenon (e.g. team learning) under study (Johns, 2006). I
will incorporate contextual details to describe conditions under which teams engage in
learning.
In some ways, team learning constructs also represent important emergent
processes within an Input-Process-Output model (Marks, Mathieu, & Zaccaro, 2001). An
emergent state is a dynamic construct that can define cognitive, affective or motivational
aspects of team members and it affects team processes and the ultimate team outcome.
3
Since, team learning essentially represents a ‘theory-in-use’ (Argyris & Schon, 1978) it is
an emergent state.
This definition of team learning is consistent with the notion of socially shared
cognition which refers to how dyads, groups and large collectives create and utilize
knowledge (Thompson & Fine, 1999). Thompson and Fine (1999) suggest that the word
‘shared’ in “shared cognition” could mean one of three possibilities. First, it could mean
“held in common” that includes overlapping cognitive representations of task
requirements and role responsibilities. Second, shared could mean “divided up into
portions” as relating to dividing up responsibility for different information. The third
possibility relates to the notion of consensus or acceptance as in viewing things and tasks
from another person’s perspective. This paper focuses on clarifying the first two
possibilities and hence will not explore the third concept of shared meaning. The two
types of sharing represent two conceptualizations of collective learning in teams.
Shared mental models (SMM) are defined as a mental representation of
knowledge regarding key components of a team’s environment that is shared amongst
team members (Mohammed, Klimoski, & Rentsch, 2000). They represent a “held in
common” type of knowledge sharing. A mental model is a cognitive structure or a
network of associations between concepts in an individual’s mind. Thus, a degree of
sharing exists amongst team members on knowledge or beliefs about the team’s
environment (Klimoski & Mohammed, 1994). A general assumption underlying SMM is
that team effectiveness will improve if the team members have an adequate shared
understanding. This shared understanding could include understanding about the task and
the equipment (task model), or awareness of team members characteristics including
4
knowledge and beliefs about appropriate or effective processes (team model) (Cooke et
al., 2003).
A transactive memory system (TMS) has been defined as a combination of an
individual’s knowledge and a shared awareness of who knows what (Austin, 2003;
Wegner, 1987). This represents a “divided up into portions” type of knowledge sharing.
TMS was initially proposed to explain the knowledge residing amongst intimate couples
and family members when they are able to bring together disparate knowledge to solve a
problem. This means that even though the solution to any issue at hand may not be
readily available, family members know how to come together and develop a response.
Wegner (1987) explains how family members develop such a thought process.
If we ask a question of a person who is a well-integrated part of a
transactive memory network, this person often is able to answer
(after consultation with other network members, of course) with
information well beyond his or her own internal storage. Asking
any member of a family a question about the family's summer
vacation, for example, can prompt the retrieval of several
members' accounts of the experience. The success we have in
retrieving certain items depends on the degree to which the person
we begin with has location information about the items we label.
Even if we ask the person to retrieve an item with an obscure label,
however, the person may be able to help us enter the storage
system. Asking Bud how much the family paid for gasoline in
Orlando, for instance, may lead him to quiz Dad-who generally
knows about car-related items (p.190).
TMS has recently been extended to the work group level as it is a cooperative
division of labor for learning, remembering and communicating relevant knowledge
(Moreland & Myaskovsky, 2000). When team members correctly identify the experts and
delegate tasks based on an individual member’s expertise, they perform better
(Hollingshead, 2000). Hence, team performance may depend on whether the group can
correctly recognize and utilize the knowledge of its members (Brandon & Hollingshead,
5
2004). The TMS definition includes two parts: a) a combination of individual knowledge
and b) interpersonal awareness of others’ knowledge.
Despite growing literature on SMM and TMS, there remains some confusion
regarding the actual conceptualization of these constructs, with proponents of TMS and
SMM suggesting that the other is a part of their construct (Austin, 2003; Mohammed,
Klimoski, & Rentsch, 2000). Specifically, Austin (2003) defined TMS as “a team mental
model about the distribution of knowledge within the group” (p. 867). Similarly, some
researchers have suggested that TMS is a Shared Mental Model about importance of who
knows what given the roles distribution in a team (Cooke, Salas, Cannon-Bowers, &
Stout, 2000; Mohammed & Dumville, 2001). While the exact nature of the relationship
between TMS and SMM is not very clear, it becomes important to clarify the unique
contribution of these constructs in particular (Mohammed & Dumville, 2001). It appears
that there is an obvious overlap between these two constructs even though they have been
conceptualized differently. Not surprisingly, researchers have a hard time defining the
term shared cognition and what is being shared (Cannon-Bowers & Salas, 2001;
Mohammed, Klimoski, & Rentsch, 2000).
To the best of my knowledge, the constructs of TMS and SMM have not been
studied together, except in a study by Ellis (2006) who indicated that lack of SMM and
TMS helps explain why teams perform poorly under acute stress. In this paper, I expand
on his conceptualization that SMM has an integrative function as it represents the
common elements whereas TMS has a differentiating function as it emphasizes
distribution of knowledge held by individual members. An integrative function captures
6
the portion of the knowledge that is universally accepted while the differentiating
function builds on the knowledge that is unique to each member in a team.
One way in which SMM and TMS are related can be understood via information
processing theory as it focuses on the cognitive processes inside a group. According to
this theory, information processing has been defined as the degree to which information,
ideas, or cognitive processes are shared and how this sharing of information affects both
individual and group-level outcomes (Hinsz, Tindale, & Vollrath, 1997). Every group
member has separate, independent memory structures. Assuming group members engage
in information sharing, they have access to one another’s memory, effectively expanding
the storage and retrieval capacity of any particular individual. Within this perspective,
SMM seems helpful in explaining how the common information is shared amongst team
members while TMS helps explain how teams collectively encode, store and retrieve
unique differentiated information. Thus, studied together they may better elucidate how
groups act as information processors and succeed in delivering results beyond any single
individual’s capacity.
My primary research questions in this dissertation are 1) How are TMS and SMM
related to each other? I plan to investigate whether these concepts are indeed distinct. In
addition, I ask 2) How do TMS and SMM impact team performance? Is one construct
better than the other in explaining team performance? Or do they better explain team
performance in combination with each other? This study is an extension of the previous
work on group learning phenomena and contributes to the literature by contrasting two
different group learning mechanisms.
7
CHAPTER 2: LITERATURE REVIEW
Theoretical framework
Learning happens at multiple levels. At times we learn independently, but most of
the time we learn by interaction with others (Bandura, 1977). These interactions can
happen at home, in the workplace, in our own network of peers and friends, and from
routines embedded in organizations (Borgatti & Cross, 2003; Gersick & Hackman, 1990;
Reagans & Zuckerman, 2001; Seger, 1994; Webb, 1982). One of the perspectives on
team learning in recent years has focused largely on SMM or TMS (Edmondson, Dillon,
& Roloff, 2008). Despite burgeoning literature on these topics, there has been only one
study that explicitly included both these concepts and clarified the nature of their
relationship (Ellis, 2006). This study used information processing theory as a framework
to suggest that SMM is about integration of teammate’s perceptions while TMS captured
differences amongst team members’ roles and responsibilities.
However, the study involved undergraduate students engaged in a computer-based
simulation of defending enemy attacks, and this makes it difficult to generalize the results
to an organizational setting. In this case, team members had clearly defined goals and
distinct expertise. While the laboratory task of keeping simulated enemy aircrafts out of
your home zone is not easy, in reality, many teams in organizational settings engage in
tasks that are much more cognitively challenging. Moreover, team members often have
overlapping expertise, lack a clearly defined role, a prior relationship with each other, and
may be required to work in future together. The peculiarities of organizational teams
make it useful to replicate in realistic settings. Additional study will help to uncover
whether these two constructs are indeed unique and help increase the external validity of
prior findings.
8
In the following sections, first, I expand on the nature of the relationship between
TMS and SMM using the information processing framework. Second, I review the
existing empirical research on SMM and TMS, as well as their impact on team
performance. Finally, I highlight the potential impact of context on the present study.
Information Processing Theory
Information processing in teams has been defined as “the degree to which
information, ideas, or cognitive processes are shared, and are being shared among group
members and how this sharing of information affects both individual and group- level
outcomes” (Hinsz, Tindale, & Vollrath, 1997, p. 53). As teams are performing
increasingly cognitive tasks, team performance relies on team members’ information
processing capabilities. Team members must attend to information in order to process it.
The information is then structured and interpreted by the process of encoding, followed
by a storage process for retrieval from memory, when necessary.
The information processing framework has been helpful in explaining team
performance in organizations. In a longitudinal study of 98 Research and Development
groups, the fit between task technology’s nonroutineness and the information processing
needs of the team helped explain team performance as far out as one year later (Keller,
1994). According to Keller (1994), information processing theory explains top
management team flexibility in strategic decision making in terms of the team’s ability to
better analyze and access information. In another cross level study, Thomas and
McDaniel (1990) suggested that a better information processing structure (low
formalization and high interaction between team members) allows top managers to better
analyze environmental factors and provides a sense of control over the strategies to be
9
followed. A CEO having lower information processing ability might evaluate a particular
situation to be threatening while another CEO with a more accurate and well-developed
information processing ability may scout for opportunities in a similar scenario. Thus,
information processing helps in sense-making and better interpretation of the
environment and could help extend our understanding of organizational learning and
change (Huber, 1991; Thomas & McDaniel Jr, 1990). In summary, information
processing theory highlights the cognitive processes by which top managers and teams
engage in learning. Thompson and Fine (1999) came to a similar conclusion in their
review of how groups develop shared meaning. According to them, groups develop
shared meaning across three main dimensions: affect, behavior and cognition.
Affective experiences include identification with a group due to reduction of
individuals’ self-identity and creation of social identity by having group-level goals
(Hogg & Terry, 2000; Tajfel & Turner, 1986). Such experiences enhance group cohesion
and group norms. Behavioral consequences include coordination and creation of products
and group decisions that are not identifiable to any individual level processes. The
cognitive dimension takes the form of TMS and SMM.
According to Thompson and Fine (1999), both TMS and SMM are manifestations
of an information-processing model. Consistent with the information-processing
approach, group members have separate and independent memory structures located
within each individual member. Moreover, group members have access to one another’s
memory if it is shared, thus effectively expanding storage and retrieval to any individual.
The notion that SMM and TMS have a lot in common as they both are concerned with
cognitive structures in teams has been reached by other researchers (cf.Mohammed &
10
Dumville, 2001). Hence, I explore the relationship between SMM and TMS using the
information-processing framework.
Shared Mental Models
The term mental model refers to a symbolic representation of a system and its
expected behavior (Johnson-Laird, 1983). According to Johnson-Laird (1983), human
beings have an innate tendency to develop and use mental models because effective
action requires an understanding of the system within which one is located. Thus people
define and enact appropriate behavior by using their evolving knowledge of the system
into a model that enables them to describe, explain and predict consequences of behavior
(Rouse & Morris, 1986). This emphasis on prediction has led researchers to explore the
usefulness of mental models primarily on performance.
Recently, the notion of a mental model was extended by researchers to account
for performance differences between teams, and SMM has been defined as an organized
understanding of relevant knowledge that is shared by team members (Klimoski &
Mohammed, 1994). As the notion of SMM was developed to explain difference in team
performance, there is an inherent assumption that SMM is an antecedent to effective team
performance (Mohammed, Klimoski, & Rentsch, 2000). SMM could help explain how
team members use similar knowledge to guide their (coordinated) behavior in effective
teams. The predictive nature of the construct could help as an indicator of a team’s
‘readiness’ to take on a particular task (Cannon-Bowers & Salas, 2001). This would lead
managers to diagnose a team’s problems and provide insight on how to solve them.
Despite the potential usefulness of the construct, the empirical evidence to support the
value of SMM is rather weak. As Cannon-Bowers and Salas (2001) note, “The problem
11
seems to be that researchers have interpreted the shared cognition label to mean so many
different things, that we are not sure that any two authors mean the same thing when they
use it” (p.196). They further highlight that researchers have used over twenty labels to
describe various types of shared cognition (e.g. collective cognition, team knowledge,
team mental models, shared knowledge, transactive memory, shared mental models).
In an early attempt to clarify the confusion regarding what is meant by shared
cognition, Cannon-Bowers, Salas and Converse (1993) proposed that there are
conceptually four distinct types of SMM. The first includes knowledge about the
equipment and tools used by the team. The second includes knowledge about task
procedures, strategies and environmental cues that impact the task. Third includes
knowledge about teammate characteristics such as preferences, habits and expertise. The
fourth component includes knowledge about team roles and team interaction patterns.
While researchers have theorized that multiple distinct types of mental models exist
(Rouse, Cannon-Bowers, & Salas, 1992), few have actually investigated multiple SMMs
and established the discriminant validity of this typology.
In some ways, the four types of models described above can be seen as reflecting
two major domains: a) Task related aspects (e.g. knowledge of equipment/technology and
job/task models) and b) Team related aspects (e.g. knowledge about team interaction and
knowledge about teammates). Consequently, most researchers have re-classified SMM
into two broad dimensions: Task-mental model and Team-mental model (Klimoski &
Mohammed, 1994, p. 432; Rentsch & Hall, 1994). Mathieu, Goodwin, Heffner, Salas and
Cannon-Bowers (2000) conceptually and empirically distinguished between task and
team-based mental models using a sample of undergraduate dyads. Further, their results
12
suggested that task mental models had no direct impact on team performance, whereas
team mental models had a positive relationship with performance.
In this paper, I focus on team-mental models rather than task models as they are
more likely to be relevant in teams that are engaged in a variety of tasks. The equipment
and task mental models are context specific, and hence harder to study in teams engaged
in different types of tasks that are not really comparable. The nature of teamwork and the
team-mental model is expected to be universally relevant compared to the task-mental
model as empirical results have been more supportive for studying team mental models
over task mental models (Mathieu, Goodwin, Heffner, Salas, & Cannon-Bowers, 2000).
Henceforth, I will use TMM to represent team-mental model. In addition, the content of
task mental model as a representation of “who does what” is much closer to the concept
of role differentiation, an issue better addressed by the TMS literature, and to be
described later.
Further, TMM involves team members’ knowledge of each other. In other words,
conceptually TMM involves shared knowledge of each other’s preferences, strengths,
weaknesses and tendencies in order to maximize performance. TMM should benefit team
performance by helping team members to compensate for each other, predict each other’s
action and provide information before being asked. Moreover, TMM is team specific and
will only hold when team membership remains relatively stable. The above clarification
responds to one of the questions posed by Cannon-Bowers and Salas (2001) that it is
critical to clarify ‘what is shared?’
Recently, Ellis (2006) examined the role of TMM and TMS in the relationship
between acute stress and team performance in a study involving 97 teams running a PC
13
based simulation task. Among other results, he found that both TMM similarity as well as
accuracy was positively related to team performance. TMM similarity refers to common
interpretation of their situation while accuracy refers to quality. The difference between
similarity and accuracy is marked by Mathieu et al. (2000) comment, ‘Similarity does not
equal quality- and teammates may share a common vision of their situation yet be wrong
about the circumstances that they are confronting’ (p.281). Ellis (2006) operationalized
TMM as the team interaction mental model, focusing on team-related aspects of a
situation. Support for a positive relationship between TMM and team performance has
also been reported by other researchers. For example, Marks, Zaccaro and Mathieu
(2000) found in a study of 79 three-person student teams engaged in a computer based
war simulation that TMM strongly predicted team performance under novel situations.
Moreover, the similarity in TMM amongst the team members was responsible for more
adaptive responses in novel situations.
Extending prior research, Smith-Jentsch, Mathieu and Kraiger (2005) used
measures of both task and team mental models to predict safety and efficiency of 306 air
traffic controllers across 47 airports. This appears to be the only published TMM study
using intact work teams. Contrary to expectations, task and team mental models failed to
demonstrate any linear effects on predicting air safety and efficiency (i.e. minimizing
airport delays). However, there was an interaction between task and team mental models
that better predicted air safety and efficiency. Their results suggested that the highest
tower efficiency and safety rates were evident when Air Traffic Controllers exhibited
high team and task mental models. Moreover, it was better to have low levels of both
team and task mental models than to have a high team and a low task mental model. It
14
appears that efforts to demonstrate linear effects of TMM on team performance have
yielded mixed and somewhat equivocal results. Hence, there is a need to study more
complex patterns of relationships to demonstrate the impact of TMM on team
performance, especially in real work teams.
One reason for the mixed results obtained between TMM and team performance
could be because of too much shared cognition. The resulting TMM might lead to
‘groupthink’, and the inability to incorporate external viewpoints might lead to failures
(Janis, 1972). Groupthink refers to a deterioration of mental efficiency, reality testing,
and moral judgment that results from in-group pressures. Groupthink has been blamed for
such decision-making fiascoes as the Bay of Pigs invasion, the escalation of the Vietnam
conflict, the Watergate cover-up, and the Challenger disaster as well as for flawed group
problem solving in business and other organizations.
Although many aspects of groupthink have been questioned, it has been
frequently invoked to explain group failures. One idea that has found support in
groupthink research is that premature consensus has a negative effect on group decision
making and leads to negative outcomes (Aldag & Fuller, 1993). When teams attempt to
learn or solve problems, often team members have divergent solutions to problems that
the team faces. Discussing these contrasting ideas is crucial to problem solving in groups
and has been considered constructive (Tjosvold, 1985). Lack of alternative viewpoints
might result in a failure to discuss critically relevant information that is not already
shared (Stasser & Titus, 1987).
In fact, when groups are comprised of familiar members, they are less likely to
bring out new or unshared information to solve a task that requires pooling of information
15
(Stasser & Titus, 1987). Such high TMM groups may already share a common frame of
reference to interpret information. It is quite possible to have high degree of consensus on
an apparently wrong way to attempt a task. As Ellis (2006) suggests, team members
under stress may have high consensus on task requirement but this may result in a bad
team performance. As Smith-Jentsch et al. (2005) observe in their study of air traffic
controllers, teams that were simultaneously high on both the task and team mental models
did not report high safety and efficiency. They speculate that a possible reason could be
that teams may overgeneralize the actions they expected from their teammates – ‘In other
words, the same implicit coordination processes that enabled these teams to perform well
under routine task conditions may have actually led to greater problems in emergency
situations’ (p.532).
Weick (1993) documented the story of fire-fighters who failed to drop their tools
because they could not comprehend an unusual direction from their team leader. In the
end, 13 fire-fighters lost their lives in the ‘Mann-Gulch disaster’. This demonstrates that a
high TMM does not necessarily result in increased performance. Another chilling disaster
attributed to air crew’s application of habitual routines resulted in the plunge of the Air
Florida flight 90 into the Potomac River shortly after take-off from Washington D.C.
(Gersick & Hackman, 1990). Thus, there are empirical and practical reasons as to why
high TMM may not essentially be a good thing.
More team interaction could be beneficial as team members engage in frequent
communication, build strong norms and increase team cohesiveness. Team members will
be able to build a social identity by having an over-arching goal (Tajfel, 1982). Such
activities will lead to a higher collective team identity that improves team performance
16
(Van der Vegt & Bunderson, 2005). A higher collective identity assumes that members
will have higher expectations from their teammates as they will have a more accurate
idea of teammate’s beliefs, abilities and preferences. Such activities are likely to increase
members’ expectations from their teammates and also improve team performance as they
come to know about each other’s strengths. However, if there is too much time spent in
understanding team members, the individual members may get too involved in
maintaining social relationships and less in their own task related activities. This is
especially true when team members have homogenous demographics and belong to a
close-knit social network (Thomas-Hunt, Ogden, & Neale, 2003). Thus, a very high level
of TMM will be undesirable as it could lead to disasters due to failure to engage in
meaningful information processing. Thus, I propose that there is an optimum level of
TMM for performance:
Hypothesis 1: The relationship between a team’s TMM and team performance will be
curvilinear (∩-shaped) such that both high and low agreement on TMM will exhibit low
levels of team performance.
Transactive Memory Systems
A TMS focuses on who knows what and attempts to capture the uniqueness of
information in a team. Based on their studies of learning in successful teams, Moreland
and colleagues (Liang, Moreland, & Argote, 1995; Moreland & Myaskovsky, 2000)
suggested that TMS has three distinct components – Specialization, Credibility and
Coordination. People in an interpersonal relationship often develop a specialized division
of labor with respect to encoding, storage and retrieval of information from different
sources. Thus, each member in the relationship develops an expertise in some areas but
not all. Other members expect members to be able to process and possess expertise in
17
specific domains. This process leads to reduction in overlapping individual knowledge
while improving information processing efficiency within teams. The Specialization
component refers to the level of knowledge differentiation within the team. Credibility
refers to team members’ beliefs about the accuracy of other members’ knowledge.
Coordination refers to team members’ ability to work together efficiently.
In one of the earliest studies indicative of TMS, Wegner, Erber and Raymond
(1991) studied memory performance of 118 individuals who had been in close dating
relationships for at least 3 months. For a memory task performed by pairs, some subjects
were paired with their partners and some were paired with an opposite-sex partner from
another couple. For some pairs, a memory structure was assigned (e.g., 1 partner should
remember food items, another should remember history items, etc.), whereas for others
no structure was mentioned. The pairs were asked to study together but not allowed to
engage in verbal communication. Memory recall was tested subsequently in individuals.
Memory performance of the natural pairs was better than that of impromptu pairs without
assigned structure, whereas the performance of natural pairs was inferior to that of
impromptu pairs when structure was assigned. This implies that dating pairs had
developed their own implicit memory structure of how to divide the task. Adding external
structure and directions resulted in interference with this TMS.
This result supports Wegner’s (1987) assertion that people often try to improve
their limited memories with external memory aids. These external aids may include either
taking notes, maintaining diaries or reliance on other people. People often turn to each
other for recalling information when they do not trust their own memory or have trouble
recalling information. Hence, it appears that TMS develops in natural groups as an
18
automatic process. This facet has spurred researchers to investigate whether teams engage
in similar memory organization and retrieval as observed in intimate couples.
Liang, Moreland and Argote (1995) in an experimental setting studied students
instructed to assemble radio kits in groups. Students trained together were better able to
assemble radios than those trained individually. Groups whose members were trained
together recalled more about how to assemble radios and made fewer assembly errors
than did groups whose members were trained apart. Later, they investigated the
videotapes of the students and coded behaviors suggesting the development of TMS. The
measures of TMS were derived from judges’ observations of videotaped teams who were
assembling the radio kits. The researchers derived observation measures for member’s
knowledge (specialization), members’ beliefs about the reliability of other’s knowledge
(credibility) and effective knowledge processing (coordination). Once these TMS
behaviors were accounted for, there were no differences between groups who were
trained together versus those trained individually. Hence, they inferred that TMS is
responsible for improved team performance. However, there could be other reasons why
these teams performed well apart from a well-developed TMS such as training and
communication.
In order to rule out alternative reasons for improved performance, Hollingshead
(1998) investigated TMS in a series of laboratory experiments. In the first experiment,
intimate couples who worked face to face performed better on a knowledge-pooling task
when compared to strangers who worked face to face or intimate couples who worked via
a computer conferencing system. Additional analyses indicated that intimate couples
when interacting face to face were better at determining when a partner’s answers were
19
correct even if only one member knew the answer prior to discussion. In the second
experiment, intimate couples scored significantly better on a knowledge task when they
had access to either nonverbal or paralinguistic communication cues than when they had
access to neither. Taken together, the results indicate that communication could be
important in the knowledge retrieval aspect of TMS.
In order to untangle communication from TMS, Hollingshead (2000) designed
two experiments to test if communication could be responsible for better group
performance, a result generally attributed to development of TMS. Sixty three, threemember teams were trained differently (e.g. individuals trained apart, individuals with
feedback about others and individuals trained together) to perform the task of assembling
radios. Groups whose members were trained apart with no chance to communicate with
each other prior to the experiment performed equally well as the groups that trained
together and whose group members were given feedback about the skills other members
had. Both the above groups did much better than groups whose members were trained
apart and had no knowledge of their members’ skills. Training techniques did not explain
differences in team performance. These set of studies clarified that TMS is distinct from
communication and training and is responsible for improved team performance via
knowledge of teammates’ skills.
TMS coordination refers to team members’ ability to work together efficiently
and has been considered critical for team performance. Research on dating couples
demonstrated that partners engage in coordinated action in response to tasks requiring an
effective TMS (Moreland & Myaskovsky, 2000). Similar coordination was observed
when a sample of clerical workers were asked to work with others in a laboratory setting
20
(Hollingshead, 2000). Participants were able to learn and recall more information when
partners had different work-related expertise. The results were reversed when they
worked with partners having similar expertise. These results suggest that coordination is
a key part of TMS as members work on their own expertise area and rely on coordination
from their partners to recall and combine information on areas different from their own.
TMS credibility refers to team members’ beliefs about the accuracy of other
members’ knowledge and provides evidence that group members trust each other’s
expertise. This dimension has also been referred to as accuracy (Austin, 2003). In a study
of student groups attempting to build AM radio-sets, groups who trained together
developed more accurate TMS than those who were trained apart (Moreland &
Myaskovsky, 2000). Researchers from other areas have also highlighted the importance
of credible recognition of expertise within teams. In a study of loan officers asked to
determine the bankruptcy chances for real estate firms, the person identified as the expert
in the team performed as well as the entire group suggesting that accurate identification
of expertise is a crucial measure of TMS (Libby, Trotman, & Zimmer, 1987). This is
consistent with Wegner’s original conceptualization that an individual can rely on other
members with more confidence when he/she has a credible source of information.
Recently, TMS has been investigated in field settings and has been shown to
explain superior team performance. Austin (2003) examined the relationships between
TMS and performance amongst 27 teams in a large apparel and sporting goods company.
Group TMS was measured as a combination of knowledge, knowledge specialization,
transactive memory consensus and transactive memory accuracy. Eleven dimensions of
skills were identified as relevant for the sample. Individuals were later asked to identify
21
the expertise level of other group-mates on these eleven skills. The individual expertise
scores were aggregated to a group score for calculating combination of knowledge.
Transactive memory consensus was defined as the extent to which group members agree
about who has the knowledge. Transactive memory accuracy involves the extent to which
individuals identified by others in the group as possessing particular knowledge actually
possessed it. According to Austin (2003), TMS is positively related to objective team
performance, external evaluations, and internal evaluations. The results proved TMS
could be applied in field settings but required direct observations and identification of
skill sets in each team that can be cumbersome. When organizations have teams with
widely different expertise it becomes increasingly difficult and time–consuming to apply
this technique of measuring TMS.
While experimental settings did provide support for the conceptualization of
TMS, it was hard to translate the measures to field settings. The measures developed for
laboratory settings (e.g. assembling radio sets) were constructed for a setting where tasks
are clear and do not differ across groups, a rarity in most organizational settings. Partly,
this was responsible for fragmented research and lower interest in its application to
organizational settings. This hurdle has been solved largely by the development of a
survey instrument validated in the field (Lewis, 2003). Lewis developed and tested a 15
item measure in a field sample of 27 teams from technology companies. Her scale
comprises the three sub-scales of specialization, credibility and coordination. The indirect
method of knowledge observation based on surveys seems as valid as direct observations
and seems to have provided a common method of measuring TMS.
22
Empirical evidence is broadly supportive of the notion that TMS impacts team
performance. It is not clear whether each of the TMS components demonstrates a positive
relationship with team performance. We need to extend the present theory by examining
the relationship of TMS components with team performance. According to Lewis (2003),
TMS is best represented as a second order factor indicated by three first order factors
(specialization, credibility and coordination). She states that “when TMS exists, it causes
specialized knowledge, mutual trust in others’ knowledge, and smooth coordinated task
processing” (p. 591). As the TMS literature is still in its nascent stage, we need to
examine the impact of each first order latent construct (specialization, credibility and
coordination) on team performance.
First, consider specialization and its impact on team performance. When team
members with distinct roles have an overlapping knowledge amongst themselves, this
causes redundancy of information. In organizational settings with clear team roles, team
members need to decide who will share what information (Mohammed & Dumville,
2001). A TMS helps a group by reducing the redundant overlaps in knowledge and
clarifying who will remember what information. This creates specialization within teams
that aids in retrieval and accessibility of information later on. Increased specialization
makes team members more efficient in cognitive processing, as only the individual
assigned to a particular expertise attends to the relevant information and encodes it to
one’s memory. This frees up other individuals to concentrate on their tasks and improves
information processing resulting in better team performance.
Coordination has been often considered critical for team performance and
effective TMS will only come from effective coordination of teammates. In a study of 35
23
five-person teams across US and Japan, Montoya-Weiss, Massey and Song (2001)
observed that coordination was responsible for smoother work flow, lowered negative
conflict and improved team performance. Faraj and Sproull (2000) in a study of 69
software teams demonstrated that the mere presence of expertise is not enough, teams
need to coordinate expertise in order to be effective. Entin and Serfaty (1999) in a study
of teams comprising Navy officials found that adaptive coordination helps teams get over
sudden events and leads to better team performance. In some ways coordination allows
teams to engage in team learning. By effective coordination, a group is able to efficiently
divide the process of acquiring information. Every member can concentrate on updating
their memory, and when needed, on retrieving the correct information from the expert
group members. In effect, TMS coordination helps in increasing the storage capacity of
the group and makes retrieval more efficient (Hinsz, Tindale, & Vollrath, 1997).
Another crucial part of TMS deals with increasing the credibility of information.
Once group members develop their expertise they will share it with their members and
over a period of time are expected to be more accurate, by reducing errors. When group
members need to find information, they are most likely to look up to the most credible
person. Thomas-Hunt, Ogden and Neale (2003) found that individuals perceived as
experts engage in more information-seeking behaviors than non-experts. They actively
share their expertise as well as engage in seeking out unique information held by minority
members. It is quite possible that credibility is in the eye of the beholder and the
individual perceived as expert might engage in self-fulfilling prophecies and thereby raise
his performance (Eden, 1984; Rosenthal, 1994). However, it is difficult to visualize the
actual processes as to why individuals who are recognized as credible engage in more
24
information sharing. Nevertheless, as long as the accuracy of team members is increased,
credibility should result in better team performance as shown in another study by Fulmer
and Stewart (2006) where a negative relationship was observed between the least
accurate team members’ perception of the leadership role and team performance.
Formally stated, I expect that the TMS components will be positively related to team
performance.
Hypothesis 2a: TMS specialization will be positively related to team performance.
Hypothesis 2b: TMS coordination will be positively related to team performance.
Hypothesis 2c: TMS credibility will be positively related to team performance.
Relationship between TMM and TMS
Many researchers have demonstrated that transactive memory systems predict
team outcomes like accuracy on tasks (Austin, 2003), team viability and team
performance (Lewis, 2004). We also know that team knowledge whether it is shared
mental models, or transactive memory systems, has been shown to result in improved
team performance (Ilgen, Hollenbeck, Johnson, & Jundt, 2005; Lewis, 2004; Mohammed
& Dumville, 2001). It is unclear whether TMS and TMM are conceptually distinct
entities. If yes, how are they related? Are they compensatory or do they interact in
different ways?
The lack of clarity between TMS and TMM can be seen in the way researchers
have treated these two constructs in past studies. While the principal proponents of each
of these constructs have tried to conceptually distinguish between the two constructs
(Lewis, 2003; Mohammed, Klimoski, & Rentsch, 2000), some researchers have started
calling for greater integration between the two constructs (Mohammed & Dumville,
25
2001). Researchers have implicitly described TMS as a type of shared knowledge (TMM)
in groups about who knows what (Rulke & Rau, 2000). Similar assumptions have been
made when TMS was defined as a form of TMM whereby team members store in
memory who is aware of what information (Fiore, Salas, Cuevas, & Bowers, 2003). Still
there are others who have equated TMM and TMS as a form of shared understanding
within teams and investigated the impact of information distribution amongst team
members on negotiation outcomes (Peterson & Thompson, 1997). It is important to
investigate the nature of the overlap and the differences between these two constructs in
order to better understand team processes and team performance.
Irrespective of conflicting viewpoints, one thing that is certain is that these
concepts are related. One explanation behind the apparent confusion is the proliferation
of similar sounding constructs of ‘shared cognition’ and the subsequent confusion
regarding the ways to operationalize and measure them (Cannon-Bowers & Salas, 2001).
According to Klimoski & Mohammed (1994), there are over thirty labels or variations of
the term ‘team mental models’. Some of these labels include terms like group cognition,
cognitive maps of collectives, strategic consensus, collective cognitions, shared frames,
shared meaning, collective mind and social cognition. It is quite likely that such labels
have increased since 1994 creating more confusion for researchers and practitioners
looking for guidance. In order to summarize and make sense of the empirical findings, we
need to be parsimonious in labeling and provide more conceptual clarity behind what we
mean by shared cognition. Once we do so it will be relatively easier to explain the
relationship between TMM and TMS, two similar sounding dominant constructs.
26
Researchers have warned that it is important to clarify what we mean by ‘shared’.
Generally, the advice is to clarify if shared means one of the following: overlapping,
similar, compatible or complementary, or distributed (Cannon-Bowers & Salas, 2001;
Mohammed & Dumville, 2001). While it is possible that some teams will be identical in
every respect and some will be totally incompatible, most teams will fall somewhere
between these extremes. In most real work teams, it is likely that some knowledge will be
shared, some will be similar, and the rest will be distributed. Hence, the a priori definition
of shared as one of the four categories will lead to a narrower definition than the reality.
I propose that we need to have an inclusive and more nuanced definition of shared
cognition. In this study, I propose that we can meaningfully capture the idea of shared
with ‘identical’ and ‘distributed’ as opposite ends on a continuum. Therefore, identical
knowledge between members would be classified as overlapping. If members have
unique knowledge, the team would be classified as distributed. This is consistent with the
idea that the shared cognition literature has overemphasized the overlapping and
underemphasized the distributed nature of sharing (Mohammed & Dumville, 2001).
Obviously, this conceptualization assumes that TMM and TMS will overlap as evident by
the coordination sub-dimension. Since both constructs, TMM and TMS, are assessed with
indices of agreement and coordination, there is an indication that such an agreement has
developed within teams.
One way to think about the relationship between TMM and TMS can be in terms
of the distribution of knowledge in a team setting. A TMM forms when group members
encode information collectively as a cognitive representation of related items (CannonBowers, Salas, & Converse, 1993; Weick & Roberts, 1993). When these mental models
27
reflect the accurate representation of team interaction, groups interact more efficiently
and perform more effectively (Mathieu, Goodwin, Heffner, Salas, & Cannon-Bowers,
2000). Thus, TMM reflects similarity amongst team members. Overlapping knowledge
amongst team members could be redundant and may result in less than optimal
performance. In such situations with clear team roles, team members need to decide who
will share what information (Mohammed & Dumville, 2001). A TMS helps a group by
reducing the redundant overlaps in knowledge and clarifying who will remember or
contribute what information. This creates specialization within teams that aids in retrieval
and accessibility of information later. The process of retrieving the correct information
from the expert group members helps in increasing the storage capacity of the group and
makes retrieval more efficient (Hinsz, Tindale, & Vollrath, 1997).
In terms of their functions, TMM is about similarity amongst team members and
suggests an integrative function amongst team members, TMS is more about
specialization of knowledge and seems to highlight differentiation amongst the team
member’s expertise (Ellis, 2006). In any team, it is quite likely that some tasks will need
integration of knowledge while at the same time unique expertise needs to be identified,
making differentiation also critical. To sum up, it appears that both TMM and TMS need
to be present in order for effective information processing in groups.
TMS ensures that information will be held by at least one of the organizational
members and lead to improved performance. In accordance with the information
processing model, TMS highlights the necessary condition for collaborative recall:
consensus, correctness and confidence (Hinsz, 1990). Despite strong theoretical support,
empirical support has been equivocal for TMS. A higher TMS does not always translate
28
into optimal team performance. This represents a process loss where team performance is
not a simple multiple of individual performance (Steiner, 1972). TMS process loss is best
captured in the following quote:
The transactive memory concept is problematic because studies of
ad hoc laboratory groups indicate that on most tasks, such as those
requiring logic, judgment, or problem-solving, groups usually
perform above the level of the average individual but rarely reach,
let alone exceed, the level of the best member (Thompson & Fine,
1999, p. 287).
The results seem puzzling since TMS has been linked to improved performance
but sometimes results in mediocre outcomes. The question arises as to why teams fail to
leverage the expertise of all their members. However, given that TMS serves as a
knowledge differentiator amongst teammates, comparison can be made from the results
found in the literature on expertise diversity and performance. The empirical results
between expertise diversity and performance have been positive in some cases and
negative in others (Ancona & Caldwell, 1992; Pelled, 1996; Wiersema & Bantel, 1992).
It appears that when diversity increases to a very high level, it becomes very difficult to
coordinate between team members. It is possible that when expertise diversity increases,
members may become alienated from each other and an even greater effort of
coordination is required so that team members can leverage on the benefits of expertise
diversity (Faraj & Sproull, 2000). Expertise diversity refers to “differences in the
knowledge and skill domains in which members of a group are specialized as a result of
their work experience and education” (Van der Vegt & Bunderson, 2005, p. 533,
emphasis added). In some ways, the notion of expertise diversity is remarkably close to
the concept of TMS as both focus on the distribution of knowledge in teams. So, we can
apply this analogy in the present paper.
29
In her dissertation, Rau (2001) found that expertise diversity explained high team
performance in the banking industry when team members develop TMS. However, her
study looked at only the successful teams and overlooked the unsuccessful ones. This
opens up the possibility that expertise diversity may have a different relationship with
team performance when TMS is low or not so well developed. Fortunately, a study by
Van der Vegt and Bunderson (2005) provides some clues to this puzzle. In their study,
expertise diversity was linked as an antecedent to team learning in a study of
multinational teams working in an oil and gas company. They examined the relationship
of expertise diversity and team performance under varying levels of collective team
identification. When team identification was low, expertise diversity was negatively
related to team performance. Conversely, with high team identification expertise diversity
was positively related to team performance.
Gruenfeld and colleagues (Gruenfeld, Mannix, Williams, & Neale, 1996;
Gruenfeld, Martorana, & Fan, 2000) in their research suggest that familiar group
members generally outperform groups of strangers when there is unique information with
each member. However, groups of strangers outperform familiar groups only when they
engage in information sharing. In some ways, TMS explains why group members do not
share unique information with team mates and only share common information for the
most part (Stasser, Taylor, & Hanna, 1989; Stasser & Titus, 1987). This may be because
team members have no idea about the possession of unique information by other
members, and they end up discussing what is collectively known. TMS will only expand
and be beneficial for the team when teammates discuss unique information.
30
When strangers do engage in mutual communication and figure out mutual
expertise, it often brings out unique knowledge. This is especially critical in tasks where a
‘hidden profile’ exists (Stasser & Stewart, 1992), a situation where team members
possess knowledge differentially and all information must be shared in order to find the
correct solution. An example includes solving a murder mystery in which the murder
suspect is correctly identified only when all participants discuss the clues available to
them. Since team members fail to share all the available information, unique information
possessed by strangers often fails to come out in the open. This situation can be
somewhat mitigated when members are motivated to complete the task and have been
informed that a clear solution will exist only if they pool their knowledge together. It is
clear that team diversity impacts how much team members will contribute to information
sharing (Thomas-Hunt, Ogden, & Neale, 2003).
Sometimes teams get motivated, often for emotional reasons, to overcome these
tendencies to avoid information sharing and engage in efforts to build a collective team
identity. According to Van der Vegt and Bunderson (2005), collective team identification
motivates team members to interact more with their teammates and re-categorize their
goals. Consistent with self-categorization theory, it is assumed that re-categorization can
mitigate the adverse effects of subgroups and create a supra-individual goal.
Consequently, evaluations of other dissimilar group members become more positive and
team members start counting others as in-group rather than out-group. When team
members engage in an integration process, it builds the shared code and language
necessary for teams to engage in knowledge creation (Nahapiet & Ghoshal, 1998).
Integration is the “process of developing shared understanding among individuals and of
31
taking coordinated action through mutual adjustment” (Crossan, Lane, & White, 1999, p.
525).
Faraj and Sproull (2000) observed that the mere presence of expertise diversity
was not enough in high performing software teams. In fact when expertise diversity
increased, team members had to engage in greater coordination so that they could
leverage the benefits of expertise diversity. I expect that a fostering environment that
creates collective identity will increase expectations of support from teammates resulting
in an increased effect of TMM on teamwork. Hence, the best team performance is
expected to occur when team members are able to balance their requirements for
encoding knowledge amongst the various team mates by building a great TMS as well as
by their ability to bring the required knowledge together through the commonality of
TMM. Great teams should be able to distribute knowledge amongst team members by
differentiating as well as integrating the collective knowledge.
On the other hand, when members are too specialized in their knowledge it may
create an isolated environment for team members. There may not be any expectations
from other team members to help, and it might result in a bunch of experts thrown
together without any shared understanding of what is required of them. This seems like a
sure recipe for disaster. Team members may stop asking for or offering information when
it is most desired simply because they may not know whom to ask. A case in point is
when the technicians building the mirror of the Hubble Telescope did not seek out the
help of designers. Being under time pressure, the technicians and designers stopped
working together, and it resulted in the production of a sub-optimal telescope. This was a
financial as well as an image loss for both the Perkin-Elmer Corporation and NASA (see
32
Ellis, 2006). Consequently, among the Perkin-Elmer team members’ TMM about team
expectations declined and the resulting team performance was lower than expected. Thus,
TMS alone may not help team performance if TMM is low.
Hypothesis 3: TMM moderates the relationship between TMS and team performance; the
relationship is negative when TMM is low but positive when TMM is high.
Context
One way in which organizational work teams differ from typical undergraduate
teams engaged in artificial tasks (e.g. playing computer games etc.) is by the presence of
different contextual variables. Context is defined as “situational opportunities and
constraints that affect the occurrence and meaning of organizational behavior as well as
functional relationships between variables” (Johns, 2006, p. 386). As Johns (2006)
argued, we as organizational researchers do not fully appreciate the impact of context in
organizational settings. Ignoring contextual differences could be a reason why field
research may not generalize across dissimilar settings. From a managerial perspective, we
need to understand the conditions under which real work teams engage in team learning
behaviors. This will be more meaningful as we understand the boundary conditions under
which laboratory results could be applied in organizational work teams. Understanding
these conditions will help practitioners prevent and diagnose potential problems in teams
who are weak on TMM and TMS.
Context can be studied at two different levels – omnibus and discrete (Johns,
2006). Omnibus is a more broadly defined entity and can best be expressed in terms of
capturing who (occupational and demographic characteristics), what (constitutes
substantive content of the research), when (refers to the time at which research was
33
conducted), where (location of the research site), and why (rationale for collection of
research data).
In contrast, the discrete dimension of context is referred to by Johns (2006) as
“the particular contextual variables or levers that shape behavior or attitudes” (p. 391).
Discrete dimensions are expected to provide the explanatory link between the more
descriptive omnibus context and specific organizational behavior like team learning. The
various dimensions of discrete dimension include task, social and physical. Johns
suggests task context would include autonomy, uncertainty, resources, etc. Similarly
some examples of social context would be social density, social structure and direct
social influence. Physical context includes temperature, light, the building environment
and décor. For the purpose of this paper, I will limit my focus to elements capturing the
task dimension of the discrete context apart from the omnibus indicators.
Research in individual job performance has sought to capture context in terms of
situational constraints that make it difficult for employees to successfully accomplish
their tasks (Peters & O'Connor, 1980). This link between situational variation and
performance variation is implicit in several theories of performance that acknowledge
how situational conditions can influence behaviors and outcomes (Campbell & Pritchard,
1976; Schneider, 1978). This notion is consistent with the interactional perspectives of
psychology (Bandura, 1986; Schneider, 1978; Terborg, 1981) that suggest variation in
context as an explanation for variance in performance outcomes. Researchers have
generally found that performance is highest when situational opportunity exists
adequately while lower performance results because of constrained settings (Blumberg &
34
Pringle, 1982; Peters, Fisher, & O'Connor, 1982; Peters & O'Connor, 1980; Steel &
Mento, 1986; Stewart & Nandkeolyar, 2006; Stewart & Nandkeolyar, 2007).
While research supports the importance of context in individual performance,
context in the form of constraints has also been found to impact team performance
(Tesluk & Mathieu, 1999). In a study of the Pennsylvania Department of Transportation
road crews, Tesluk and Mathieu (1999) observed that crew performance was hampered
due to severe problems with their equipment, materials and work procedures. They
further report that “Road crews who experienced more significant barriers to performance
were rated by their managers as being less effective in meeting work deadlines,
overcoming problems, maintaining effort, providing quality service, and working as a
cohesive crew” (p. 210). It is crucial that teams effectively counter constraints and take
pre-emptive measures to avoid lower team performance. As management researchers, we
should be able to account for such contextual barriers to better account for team
performance.
Organizational context is a significant source of influence on group processes.
The existing research on context in teams can be reviewed by two distinct approaches
(Mowday & Sutton, 1993). The first focuses on an organizations’ impact on teams
(Hackman, 1987, 1990), the second examines the teams attempt to influence the larger
organization (Ancona & Caldwell, 1992). According to Mowday and Sutton (1993), one
way to study context is in terms of opportunities and constraints placed on the teams.
Murnighan and Conlon (1991) in a study of British string quartets illustrated how context
constrained behavior. In particular, the task of producing music influenced the resolution
of paradoxes – leadership vs. democracy, confrontation vs. compromise – within the
35
string quartets. Other examples of context influencing team processes abound (Gladstein,
1984; Katz, 1982).
Katz (1982) found that team longevity, measured as the time team members spent
working together, affects group processes. In particular, the length of time for which 50
R&D teams worked together impacted teams’ communication levels and team
performance. Among newer teams, increased time working as a team was associated with
better team performance. In contrast, amongst more mature teams increased time together
correlated with decreased performance. Katz (1982) suggested that over time, groups
become increasingly insulated from internal and external communication vital for their
performance.
In yet another study involving 100 sales teams in the communication industry,
Gladstein (1984) investigated the drivers of team member satisfaction, team members’
self-reported effectiveness and sales performance. While the traditional theories of group
effectiveness based on team inputs like task, structure and composition predicted team
member satisfaction and self-reported effectiveness, it failed to account for the actual
sales performance. She found that most of the variance in the team’s sales performance
was due to organizational contextual variables like rewards, supervisory leadership and
industry growth rate. She suggested that there is a need to better understand how
contextual factors like task demands and organizational level variables influence team
outcomes.
In a study involving 51 surgical teams across multiple hospitals, Edmondson
(1999) investigated the antecedents, processes and outcome variables associated with
team learning. Edmondson found that contextual factors (e.g. organizational support and
36
team leader coaching behavior) impact learning behaviors and team performance
outcomes through psychological safety mechanisms. Psychological safety was described
as the shared belief that the team is safe for interpersonal risk taking. Edmondson (1991)
called for more integrative research by investigating task (e.g. team structures) and social
(e.g. psychological safety) dimensions of contextual factors.
In one of the strongest evidence of contextual influence on team learning, Gibson
and Vermeulen (2003) found that demographic characteristics (subgroups within team)
fostered team learning behavior in a pharmaceutical firm. They found that a moderate
‘subgroup strength’ helped team learning. Subgroup strength was defined as the degree of
overlap across multiple demographic characteristics among a subset of team members.
Both very heterogeneous and very homogenous teams were not as effective in team
learning compared to a team with moderate differences (subgroups). Further, in teams
with moderate subgroups, other contextual factors (empowerment, knowledge
management systems) accentuated team learning behavior.
Further support for studying contextual effects on team learning comes from the
related field of team creativity and innovation. In a review article on team creativity and
innovation implementation process, West (2002) articulated that task characteristics are
an important antecedent on team creativity. He suggested that team autonomy and task
requirements of completeness, opportunities for social interaction, opportunities for
learning, and developmental possibilities lead to increased intrinsic motivation for team
members and should be taken into account.
Now, I return to the original hypotheses and discuss the potential contextual
impact on the hypothesized relationships. The first hypothesis proposes a curvilinear
37
relationship between TMM and team performance. This relationship rests on the unstated
assumption that very high agreement within team is detrimental. There are certain tasks
in which instantaneous action is required from individuals. Any wrong action could lead
to potential disaster and yet, we hardly hear of accidents in what have been called highreliability organizations (Weick, 1987). According to Weick (1987), examples of such
organizations would include air traffic control, nuclear power plants, naval air-craft
carriers etc. In such a situation, we may fail to see the curvilinear relationship because the
nature of the task requires a very high level of sensemaking to avoid accidents and
consequently a very high level of TMM. The team performance will probably not lower
if the teams have higher amount of TMM and everyone anticipates how their teammates
will be reacting. In this case, the team performance may exhibit a more linear relationship
with TMM.
In contrast, a manufacturing plant like Toyota Motors may like to avoid a very
high TMM amongst its workforce. Toyota’s organizational culture requires people to
suggest continuous improvements in production techniques, engage in reflection,
continuous learning and knowledge sharing (Dyer & Nobeoka, 2000; Liker, 2004). It is
quite likely that team performance will be best when teams display a moderate level of
TMM. A moderate level of TMM allows for flexibility in teams to suggest continuous
improvements (kaizen) which may not be possible with constrained routines expected
due to a very high TMM. Moreover, employees are motivated to engage in knowledge
sharing and deviate from expected norms as consequences of mistakes are generally not
as serious in an automotive plant as compared to a nuclear power plant control room.
38
Hence, we should be mindful of the research setting encouraging high TMM versus those
that don’t.
Of course, the relationship may depend on other contextual factors like team
interdependence. It is quite possible that teams engage in a sequential process where
minimal coordination is required in terms of inputs from other team mates. Such a
scenario will suggest that team interdependence will be much lower. Further, the team
performance will be somewhat independent of the TMM and hence the proposed
hypothesis may not be true. In this case, the relationship between TMM and team
performance may not be linear but more likely to be horizontally flat, irrespective of
TMM level. The preceding discussion highlights the importance of contextual effects to
the study.
A similar case can be made for context influencing the relationship between TMS
and team performance, as a function of task type. The nature of the task carried out by the
team is quite important. If the task is relatively simple and routine, or an additive task, a
team might be able to perform adequately even if the team members have not developed
TMS. The relatively simple nature of the task may not require a substantial level of
coordination and the expertise of different team members as part of a TMS. Hence, task
complexity could be one contextual component which may weaken the association of
TMS components with team performance. Of course, other contextual factors like a
strong reward for team support could also influence the nature of relationships by
motivating team members to work together. An external cue like a team based reward
system could aid development of TMS in teams.
39
While I am not discussing the impact of different contextual variables on the third
hypothesis which delves into the relationship between TMM and TMS, the previous
examples suggest that research settings as well as discrete contextual variables focused
on task can significantly alter results. After all, the need to really find the middle ground
between commonalities (TMM) and the differences (TMS) does not have to be important
in all teams. The proposed relationship between TMM, TMS, and team performance
depends on the nature of the task and the type of the team. Though various typologies
exist to classify teams, I will use the taxonomy developed by Sundstrom, de Meuse and
Futrell (1990). They define team types based on the degree of internal differentiation
(member heterogeneity) and external integration (linkage to organizational activities).
The team types are advice/involvement, production/service, project/development and
action/negotiation teams.
Advice/involvement teams are homogenous decision-making committees
comprised of first-line employees whose actions were loosely coupled to the working of
the organization. Production/Service teams are front-line employees that provide products
or services and whose activities are highly integral to daily operation of the organization.
Project/development teams are a homogenous group of white-collar professionals
collaborating on one-of-a-kind projects whose operations are weakly aligned to the
organizations’ routine activities. Advice/Negotiation teams are groups of highly
specialized individuals that engage in relatively brief real-time performance events
requiring improvisation in unpredictable circumstances and whose activities are highly
integrated with the rest of the organization.
40
It is clear from the above summary that context could be quite important for team
learning and influence the relationships between group processes and team outcomes. I
include both omnibus (team demographics) as well as discrete features (team identity,
team interdependence, reward system and organizational support) of contextual features.
Moreover, In order to detect and appreciate contextual effects, I include multiple
dependent variables (Johns, 2006). Multiple dependent variables help uncover the impact
of contextual effects as some variables are less sensitive to context than others. Hence,
apart from measuring team performance as an outcome of TMM and TMS, I also include
other theoretically relevant dependent variables. Such variables include performance
related outcomes such as team member satisfaction and team viability (Hackman, 1987),
team learning (Edmondson, 1999) and team creativity (Amabile, Conti, Coon, Lazenby,
& Herron, 1996).
In summary, the contextual variables perform two important functions in this
study. First, they add details that lead to a richer understanding of subjects and their work
settings. This is consistent with the call by Johns (2006) mentioned above. Second, these
contextual variables specify the boundary conditions of the proposed model. The
importance of TMM and TMS is probably more relevant due to internal or external
factors affecting the team. Internal factors include inherently complex tasks,
interdependent team members etc. External cues include team rewards to motivate team
members working together and teams engaged in repeated interactions (e.g. problemsolving project teams or production/service teams). Figure 1 summarizes the impact of
contextual variables on the TMS, TMM and team effectiveness. Figure 2 depicts the
hypothesized relationships as mentioned in the previous section.
Contextual Variables
Team effectiveness
•
•
•
•
•
•
Team tenure
Team size
Country of origin
Task interdependence
Team reward
Organizational support
•
•
•
•
•
Team performance
Team viability
Team satisfaction
Team learning
Team creativity
Emergent processes
•
•
Team mental models
Transactive memory
systems
Figure 1. Team-level conceptual model describing relationships between processes, context and dependent variables.
41
Team Mental Models
H1
Transactive Memory
Systems
H3
TMS
Specialization
Team Performance
H2a
TMS
Coordination
TMS
Credibility
H2b
Team tenure
Task
interdependence
Team size
Org support
Country
Team reward
H2c
.
Figure 2. Hypothesized relationship between TMM, TMS and team performance
42
43
CHAPTER 3: METHOD
Procedure
I administered an online version of the survey to team members and their team
managers. The survey was expected to take no longer than 15 minutes for the team
members and had a separate shorter section for the team managers. All the survey items
are detailed in Appendix A. Team managers provided measures on team effectiveness
dimensions (team performance, team learning and team creativity). Team members were
asked to provide information on the rest of the variables including team learning
measures (TMS and TMM). This helped minimize the common-method bias by having
separate sources for independent variables and dependent variables. An informed consent
document preceded the online survey and explained the purpose of the study to survey
respondents (Appendix B).
I enlisted the help of sponsoring organization managers in administering the
survey. Typically, senior managers in the sponsoring organizations forwarded my
requests to their teams (team members and respective team managers); this way the
anonymity and confidentiality of the survey respondents were maintained. Given that
team based research requires a substantial number of teams to do a meaningful analysis
(Stewart, 2006), I ended up surveying teams from multiple organizations as no one
particular organization was sufficiently large in terms of team size. This process assured
substantial variation between teams and was expected to result in a sample generalizable
across multiple settings. I recorded organizational characteristics and team demographics
to control for differences amongst teams from multiple organizations. I recruited teams
who were working in the service industry from both the US and India, and which were
part of project teams or ongoing work teams (Cohen & Bailey, 1997).
44
Sample
I approached organizations that were team-oriented and had teams working on
clearly defined tasks or projects. Since organizations differ widely in their definition of
teams, I prescreened the teams that agreed to participate in the study by employing the
four criteria as described by Guzzo and Dickson (1996): 1) Teams must be part of an
organization, 2) Teams have unique entity and clear boundaries, 3) Team members
perform interdependent tasks, and 4) Tasks affects coworkers or customers.
In summary, 363 members from 46 teams participated in my survey. The team
size ranged from 3 to 32 members and averaged 11 members. Of the 363 total team
members surveyed, 290 provided usable responses. Of the usable response at the team
level, the final response rate was 58% within teams. The actual number of respondents
within a team ranged from a minimum of two (2) to a maximum of twenty-five (25).
Team size was included in the team level analyses (Hypotheses 1-3). Ratings of
individual teams by their respective team managers were available for only 41 out of the
46 teams surveyed. Hence, complete data from 41 teams were further analyzed.
Eighty-one percent of the team members had a bachelors’ degree or higher. The
average team member had been with the current organization for 6.27 years and current
team for 2.91 years (S.D. = 2.16). Average team size was 11 members (SD = 7.4) and
average number of respondents in a team was 5.50. Sixty-three percent of respondents
were male. Slightly over forty-six percent (46.3%) of sample reported their ethnicity as
Caucasian, 1.7% Hispanics, 51.7% Asians and 0.4% Hawaiian/others. The higher than
expected sample of Asians is because of respondents from the Indian Information
Technology (IT) industry.
45
My sample consists of teams from the US and India. The teams in the US were
drawn from six organizations belonging to the Aeronautical, Healthcare, Education,
Engineering, Information Technology (IT) and Retail industries. The teams from India
were drawn from four organizations, three from the IT Industry and one from a finance
company. More details about the teams, their country of origin and the industry are
available in Table 1.
Aggregation Issues
Various methods of computing within group agreement exist and the exact
statistic used to compute agreement depends on the underlying theoretical issue (Bliese,
2000). TMM fits the definition of a dispersion model of composition “because it refers to
the variability within a group and a variance statistic is indexing an attribute of a group as
opposed to an attribute of any individual-level response” (Chan, 1998, p.239). Statistics
used to capture TMM, a team level agreement, could include rwg (James, Demaree, &
Wolf, 1993), ICC (Shrout & Fleiss, 1979), Average Deviation (Burke & Dunlap, 2002) or
Standard deviation (Schmidt & Hunter, 1989).
Similar to the TMM, all constructs measured in the survey were conceptualized at
the team level though the responses were measured at the individual level. To justify the
aggregation to the team level, I calculated within-group agreement by computing
interrater agreement on a uniform expected variance distribution commonly abbreviated
as rwg(j) or rwg (James, Demaree, & Wolf, 1984). The rwg is calculated by comparing an
observed group variance to an expected random variance - the latter usually refers to
calculating a uniform distribution. The uniform distribution refers to the obtained
distribution when all group members provide the same number of responses for each
46
category i.e. if group members provided an equal number of 1s, 2s, 3s, 4s and 5s on a 5point scale. This kind of distribution is seldom seen in groups as team members typically
suffer from response bias e.g. team members are more likely to use 3, 4 and 5 instead of
the full range. The presence of such response bias results in reduced observed variance
and a high value for rwg. Hence, the use of rwg has been criticized (cf. Schmidt & Hunter,
1989). Inspite of the shortcomings, the magnitude of the rwg provides a measure of
within-group agreement on the variables of interest. The summary rwg for the team-level
variables are presented in Table 3. Average interrater agreement was above the 0.70
benchmark proposed by James et al. (1984) for all variables except TMS coordination.
As an alternative to rwg, I also calculated two intraclass correlations – ICC(1) and
ICC(2) and conducted an F-test for the ICC(1). ICC(1) indicates the extent of agreement
among ratings from members of the same team whereas ICC(2) indicates whether teams
can be differentiated on the variables of interest (Bliese, 2000). ICC(1) has been
interpreted as an index of interrater reliability, the extent to which the raters are
substitutable, and hence a suitable measure for aggregation (James, 1982). ICC(1) was
calculated using the one-way random-effects ANOVA model:
ICC(1) = (MSB-MSW)/[MSB+[(k-1)*MSW].
MSB is the mean square between-group, MSW is the mean square within-group,
and K is the average team size. ICC(2) provides an estimate of the reliability of the group
means and is calculated by the following formula: ICC(2) = MSB-MSW/MSB. Details of
ICC(1), ICC(2) and rwg are reported in Table 2.
F-test for ICC(1) was significant for all the variables except for team satisfaction
and team reward. However, I computed ICCs for all the variables and aggregated them
47
anyway since the constructs are theorized at the team level of aggregation. As can be seen
from the Table 2, ICC (1) and ICC(2) for TMS coordination was .10 and .40 (F=1.67,
p=.01), task interdependence was .14 and .50 (F=2.00, p=0.00), Organizational support
was .14 and .50 (F=2.00, p=0.00) and team viability was .14 and .49 (F=1.96, p=0.00).
These ICC values indicate that we don’t have reliable estimates of the group means and
are on the lower side but are comparable with values obtained in field research conducted
elsewhere (cf. Chen & Klimoski, 2003). Hence, these values of ICC should be acceptable
in the present study.
Study Variables and Measures
Control variables
Heterogeneity measures including subgroups have been indicated to impact team
learning (Gibson & Vermeulen, 2003). Hence, I recorded sex, age and ethnic background
in the survey to control for the effects of heterogeneity in teams. Moreover, I recorded the
type of task being done by the teams. The team task was classified as one of the four
types- advice/involvement, production/service, project/development and
action/negotiation teams based on the team managers’ response (Sundstrom, de Meuse, &
Futrell, 1990). Additionally, I controlled for the country of origin (0=India, 1= US). The
sample was made up of twenty-six teams from the US and fifteen teams from India. In
terms of industry, I had 14 teams from IT, 12 from Aerospace, 7 from Education, 5 from
Healthcare, and one team each from Engineering, Retail and Finance. All teams from the
IT industry were composed primarily of Indians or people of Indian origin working in the
US. I also controlled for team tenure as teams with longer tenure might have a history of
interaction amongst their team, so they may behave differently as they build up rituals
48
and norms (Katz, 1982). Team size was also recorded as a control variable since
coordinating a large number of teammates could be an important factor. These contextual
features were expected to impact the hypotheses and I measured them accordingly
(Johns, 2006).
Team Mental Model
Team-mental model refers to a shared understanding of knowledge about team
interaction and team members (Mathieu, Goodwin, Heffner, Salas, & Cannon-Bowers,
2000). Stated otherwise, team-mental model involves shared knowledge of each other’s
strengths, weaknesses and preferences. Typically, TMM has focused on aspects of the
team (e.g. teammates’ knowledge, skills, attitudes and preferences) and team interaction
(e.g. roles, communication channels, information flow and role interdependencies). One
way to operationalize this shared team knowledge has been to measure the degree of
convergence amongst teammates. The convergence can be assessed by asking team
members to respond to a structured questionnaire about their teammates and team
processes. A high degree of overlap amongst team members should suggest a high degree
of TMM. Since this was an exploratory study in a field setting to studying team mental
models, in the absence of prior research guidance I operationalized this construct in two
different ways. The first uses an average deviation method and the other a 6-item five
point Likert type scale I developed specifically for this study.
The use of agreement indices to capture mental models has been a common
practice (Mohammed, Klimoski, & Rentsch, 2000). In particular, Mathieu and colleagues
(Marks, Zaccaro, & Mathieu, 2000; Mathieu, Goodwin, Heffner, Salas, & CannonBowers, 2000) have measured TMM by assessing the overlap in concepts selected by
49
each of the team members. TMM was assessed by teammates’ individual ratings of the
relationship between various attributes. Mathieu and colleagues (Marks et al., 2000;
Mathieu et al., 2000) asked undergraduate subjects to rate the relationship between team
process on nine dimensions. In essence, each subject in every team filled up a 9X9 matrix
on team processes. It was extremely difficult to sell this idea of filling up a 9X9 matrix to
working professionals to collect a single measure. In the interest of economizing on the
time taken to complete the survey, I used a slight variation on the basic process.
I selected six out of nine team processes from the one used by Mathieu et al.
(2000) and asked the team members to rank order their relative importance. A high
convergence in the rank-ordered pattern between various team members suggests a strong
convergence on TMM. I selected the six team processes based on the existing literature
(Marks, Mathieu, & Zaccaro, 2001).
Marks et al. (2001) suggest that teams have ten process dimensions nested within
three superordinate categories: a) transition phase processes, b) action phase processes
and c) interpersonal processes. Transition phases are periods of time when teams focus
primarily on evaluation and/or planning activities to accomplish team objectives. During
the action phase, teams conduct activities leading directly to goal accomplishment.
Interpersonal processes are used to manage interpersonal relationships within teams.
Transition phase processes include mission analysis, goal specification and strategy
formulation processes and typically occur during the time period when teams reflect,
evaluate and plan for future directions. Action processes refer to monitoring progress
towards goals, systems monitoring, team monitoring and back-up behavior and
50
coordination. Interpersonal processes include conflict management, motivation and
confidence building and affect management.
Given that my approach was to make it easy to collect data from professionals
working on teams, I selected the six most important and practical team processes likely to
be generalizable across all teams. Transition processes such as goal specification and
strategy formulation and action processes such as monitoring progress towards goals,
team monitoring and backup behavior are crucial to understanding team outcomes.
Moreover, since I focused on understanding team satisfaction and team viability
outcomes, I included the interpersonal processes that occur throughout both transition and
action processes – conflict management and motivation/confidence building.
First, consider the transition phase processes. It should be noted that mission
analysis has not been included in this study since in most cases the teams’ mission,
resources and operating conditions are often dictated by senior executives and are beyond
the jurisdiction of team members. This forces team members to engage in goal
specification and strategy formulation to achieve the stated mission. Goal specification
refers to identification and prioritization of goals and subgoals towards achievement of
the teams’ mission. Teams engage in goal specification to indicate what and how much
must be accomplished within a specified time and within certain quality standards.
Strategy formulation involves how team members go about achieving their missions,
discussing expectations, relaying task-related information, prioritizing, role assignments,
and communicating plans to all team members.
Second, I reflect on the relevant action phase processes. Monitoring progress
toward goals is defined as tracking task and progress toward mission accomplishment,
51
interpreting system information in terms of what needs to be accomplished for goal
attainment, and transmitting progress to team members. Poor goal monitoring occurs
when teams drift, procrastinate, or stray off task and are unable to provide appropriate
performance feedback. Hence, it is vital that teams engage in periodic goal monitoring.
Another important action process is team monitoring and backup behavior. Assisting
team members happens by providing verbal feedback, task-related support and carrying
out and completing tasks for their teammates. Team members observe their teammates’
actions and watch for performance discrepancies. Whenever team members identify the
need to provide help, backup behavior in terms of suggestive or corrective feedback is
provided to get performance back on track.
Third, I expand on the two interpersonal processes – conflict management and
motivating/confidence building. Conflict management refers to attempts to manage
conflicts within the team. These can be preemptive or reactive. Preemptive techniques
involve establishing conditions to prevent, control or guide team conflict before it occurs.
Reactive techniques involve working through task, process and interpersonal
disagreements among team members. Managing conflict is critical as it has been
suggested to lower team performance (De Dreu & Weingart, 2003). The other critical
interpersonal process is motivating/confidence building which involves encouraging team
members to perform better or to maintain high levels of performance. Teams can enhance
working relationships and performance by boosting their teammate’s confidence levels.
In contrast, team members can engage in negative comments that can lower team
confidence and performance over time.
52
Participants were asked to rank order the importance of these six team processes
in their team. Thus, each team member has his or her own rank orders.
Team Mental Model (Average Deviation)
I calculated Average Deviation (AD) indices for the level of TMM within each
team (Burke & Dunlap, 2002; Dunlap, Burke, & Smith-Crowe, 2003). The resulting AD
indices, from the rank ordered response on six team processes, were used as independent
variables to capture TMM level in each team. Finally, I multiplied the resulting AD by (1) to reflect the direction of the construct from disagreement to agreement within team.
Team Mental Model (alternative scale)
In addition to computing Average Deviation on the six team processes, I
developed a 6-item scale to capture each of the six team processes as described in the
previous section. Sample items include “We identified and prioritized goals for task
accomplishment”, “We tracked our progress toward task accomplishment” and “We
established rules to prevent, control, or guide team conflict before it occurred and worked
out interpersonal disagreements among team members”. I averaged teammates response
within each team and then averaged items to form a single scale. The coefficient alpha
measured at the individual level was 0.83. The rwg value was 0.82. ICC(1) and ICC(2)
were 0.21 and 0.62 respectively suggesting the presence of substantial within-group
agreement and allowing aggregation to the team level. I ran a confirmatory factor
analyses to test whether the 6-item scale fits a one-factor model or not and the one-factor
model showed a great fit to the data (χ2 = 16.54, df = 9, p<.10; SRMR = .05, NNFI = .99,
IFI = .99, CFI = .99).
53
Transactive Memory Systems
Transactive Memory System focuses on who knows what and is influenced by
knowledge about the memory system of another person (Wegner, 1987). Since each team
member has a different way to store, encode and retrieve information, the knowledge
held by individuals is often unique leading to specialization in information processing.
The TMS develops when individuals develop accurate perception about their teammates
and credibility develops as team members’ form beliefs about the accuracy of other
member’s knowledge. However, just specialization and credibility can’t help develop
TMS unless they engage in transactions. These transactions often take the form of
communication and interpersonal interactions, requiring a certain level of coordination
between team members. Thus, any measure of TMS should be able to capture these three
facets – specialization, credibility and coordination (Austin, 2003; Lewis, 2003).
Therefore, absence of any of these three dimensions hampers the development of TMS
and interferes with the efficient distribution of knowledge within teammates.
To assess TMS and its subcomponents, I adapted the 15-item questionnaire
developed by Lewis (2003) for field studies. She developed and tested these items in a
laboratory sample of 124 teams, 64 MBA consulting teams and 27 work teams from
technology firms. Her scale has five questions for each of the three TMS facets specialization, credibility and coordination. Based on results of confirmatory factor
analyses, Lewis (2003) proposed that TMS is best represented as a second-order factor
indicated by three first-order factors (specialization, credibility and coordination). This
scale has an advantage of being task-independent and can be used to capture indirect
measures of TMS with a high validity. TMS is conceptualized as a shared team level
54
construct and hence teammates’ responses were aggregated to the team level. This is
because the survey items have a ‘referent shift’ since the items pertain to questions about
the team and not the individuals themselves (Chan, 1998).
Consistent with Lewis (2003), I created one composite scale to measure TMS.
However, I separately computed the three first order scales of TMS specialization, TMS
credibility and TMS coordination. In order to examine the factor structure of the TMS
variables (Specialization, Coordination and Credibility), I specified three measurement
models using LISREL 8.71 (Jöreskog & Sörbom, 1996). A single factor to capture TMS
provided a poor fit to the data (χ2 = 213.03, df =90, p<.01; SRMR = .08, NNFI = .83, IFI
= .86, CFI = .86). Both the three-factor model with Specialization, Coordination and
Credibility and second order factor with a single factor TMS indicated by three first
order factors (TMS Specialization, Coordination and Credibility) were statistically
equivalent and provided good fit to the data (χ2 = 116.57, df = 87, p<.01; SRMR = .08,
NNFI = .96, IFI = .97, CFI = .97). Because the latter two models were better fitting than
one based on a single TMS factor, I continued exploring the three-factor TMS construct
in the subsequent analyses. I also analyzed the single factor TMS construct in keeping
with prior research but expected the three-factor model to yield better insights. I have
included multiple fit indices to report the results of the Confirmatory Factor Analyses.
Table 3 details the result of the Confirmatory Factor Analyses on all study variables.
TMS Specialization
This construct was measured by the five-item scale developed by Lewis (2003).
Sample items include “each team member has specialized knowledge of some aspect of
our project”, “Different team members are responsible for expertise in different areas”
55
and “I know which team members have expertise in specific areas”. I averaged
teammates response within each team and then averaged items to form a single scale. The
coefficient alpha was 0.61. The rwg value was 0.84. ICC(1) and ICC(2) were .28 and .70
respectively, suggesting the presence of substantial within-group agreement and allowing
aggregation to the team level.
TMS Credibility
I measured TMS credibility using the five-item scale developed by Lewis (2003).
Sample items include “In most cases, I trusted that other members’ knowledge about the
project was credible”. I added the phrase “In most cases” in front of all the items. This
was done to get a global sense of credibility and not let team members get distracted by
isolated incidents. I averaged responses from teammates within each team and then
averaged items to form a single scale. The coefficient alpha was 0.77. The rwg value was
0.84. ICC(1) and ICC(2) were .42 and .81 respectively, suggesting the presence of
substantial within-group agreement and allowing aggregation to the team level.
TMS Coordination
I measured TMS coordination using the five-item scale developed by Lewis
(2003). Sample items include “We accomplished the task smoothly and efficiently” and
“our team worked together in a well-coordinated fashion”. I averaged responses from
teammates within each team and then averaged items to form a single scale. The
coefficient alpha was 0.66. The rwg value was 0.55. ICC(1) and ICC(2) were .10 and .40
respectively, suggesting lack of within-group agreement. However, in order to be
consistent with similar measures (e.g. TMS specialization and TMS credibility), I
aggregate the items into a team-level scale.
56
TMS
I measured the TMS construct by using the 15 items scale by Lewis (2003). This
scale is a composite of the three subscales described earlier in the preceding paragraphs TMS specialization, TMS credibility and TMS coordination. The coefficient alpha was
0.79. The rwg value was 0.79. ICC(1) and ICC(2) were .28 and .70 respectively,
suggesting presence of substantial within-group agreement and allowing aggregation to
the team level. As noted earlier, based on the confirmatory factor analyses, the singlefactor construct was a poor fit to the data (χ2 = 213.03, df =90, p<.01; SRMR = .08,
NNFI = .83, IFI = .86, CFI = .86).
Team Viability
Team viability measures a teams’ focus to maintain itself over time. Given that
work teams work on multiple projects over a period of time, team viability can be a vital
measure of effectiveness distinct from team performance. Team members rated their team
using the seven-item measure developed by Hackman (1988). Sample items include
“members of the team care a lot about it, and work together to make it one of the best”,
“working with team members is an energizing and uplifting experience” and “as a team,
this work group shows signs of falling apart”. Coefficient alpha was 0.86. The rwg value
was 0.88. ICC(1) and ICC(2) were relatively lower at .14 and .49 respectively, suggesting
the presence of relatively low within-group agreement. I ran a confirmatory factor
analyses to test whether the 7-item scale fits a one-factor model or not and the model
showed an excellent fit to the data (χ2 = 69.46, df = 14, p<.01; SRMR = .07, NNFI = .95,
IFI = .97, CFI = .97).
57
Team Satisfaction
Team satisfaction was measured by a three-item scale developed by Hackman
(1988). Team members rated their agreement with sample items such as “Generally
speaking I am very satisfied with the team”, “I frequently wish I could quit the team
(reverse coded)” and “I am generally satisfied with the work I do on the team”.
Coefficient alpha was acceptable at 0.73. The rwg value was 0.84. ICC(1) and ICC(2)
were .03 and .16 respectively, suggesting lack of within-group agreement to allow
aggregation to the team level. However, given that ICC(1) and ICC(2) were low while the
rwg value was reasonably high at 0.84, it seems that most teams are relatively similar in
response to team satisfaction. This might have lead to low within-group to between-group
variance and substantially lower ICC values. In keeping with the rest of the variables, I
aggregate the items into a team-level measure. The confirmatory factor model is not
reported because the 3-item scale indicated a saturated model and demonstrated perfect
fit.
Team Interdependence
I used a modified version of task interdependence items developed by Kiggundu
(1983). The five-item scale was used to assess employees’ task interdependence at an
engineering company by Van der vegt et al. (Van der Vegt, Emans, & Van de Vliert,
2001). Sample items include “I have to obtain information and advice from my
colleagues in order to complete my work”, “I have a one person job; I rarely have to
check or work with others” and “In order to complete their work, my colleagues have to
obtain information and advice from me”. The coefficient alpha for the scale was 0.74.
The rwg value was 0.74. ICC(1) and ICC(2) were .14 and .50 respectively, suggesting the
58
presence of relatively low within-group agreement to allow aggregation to the team level.
I ran a confirmatory factor analyses to test whether the 5-item scale fits a one-factor
model or not and the model showed a great fit to the data (χ2 = 13.50, df = 5, p<.01;
SRMR = .04, NNFI = .93, IFI = .97, CFI = .97).
Team Reward
I used the three-item scale developed by Denison, Hart and Kahn (1996). Sample
items include “my performance review depends upon my performance as a member of
the team, “my performance review depends upon the performance of the team and
“effective work in support of teams is critical to my advancement within the
organization.” The coefficient alpha for the scale was 0.58. The rwg value was 0.68.
ICC(1) and ICC(2) were .05 and .23 respectively, suggesting lack of within-group
agreement. However given that ICC(1) and ICC(2) were low, it appears that most teams
are relatively similar in response to team satisfaction. This might lead to lack of betweengroup variance and substantially lower ICC values. In keeping with the rest of the study
variables, I aggregate the items into a team-level measure. The confirmatory factor model
is not reported because the 3-item scale indicated a saturated model and demonstrated
perfect fit.
Organizational Support
I used the five-item scale developed by Edmondson (1999). Sample items include
“my team gets all the information it needs to do our work and plan our schedule”, “my
team is kept in dark about current developments and future plans that may affect its
work”. The coefficient alpha for the scale was 0.70. The rwg value was 0.74. ICC(1) and
ICC(2) were .14 and .50 respectively, suggesting lack of within-group agreement.
59
However given that ICC(1) and ICC(2) were low, it appears that most teams are
relatively similar in response to team satisfaction. This might lead to lack of betweengroup variance and substantially lower ICC values. Finally, I ran a confirmatory factor
analyses to test whether the 5-item scale fits a one-factor model or not and the and the
model showed excellent fit to the data (χ2 = 10.16, df = 5, p<.10; SRMR = .09, NNFI =
.96, IFI = .98, CFI = .98).
Team Performance
I adapted the six-item team performance measure developed by Kirkman and
Rosen (1999) to capture ratings of team performance. Sample items include “the team
meets or exceeds it goals”, “team completes its tasks on time” and “team is a productive
team”. I asked the team managers to rate their team. The coefficient alpha for the scale
was 0.83. I ran confirmatory factor analysis to test whether the 6-item scale fits a onefactor model. The model showed a reasonable fit to the data (χ2 = 14.52, df = 8, p<.10;
SRMR = .08, NNFI = .91, IFI = .95, CFI = .95).
Team Learning
Team learning was adapted from the five-item scale developed by Edmondson
(1999). Team managers rated their team using items such as the team “asks its internal
customers for feedback on its performance”, “relies on outdated information or ideas
(reverse scored)”, and “regularly takes time to figure out ways to improve its work
performance”. Coefficient alpha was 0.63 suggesting the scale reliability was marginal. I
ran confirmatory factor analysis to test whether the 5-item scale fits a one-factor model or
not and the model showed an acceptable fit to the data (χ2 = 5.46, df = 4, n.s.; SRMR =
.12, NNFI = .90, IFI = .97, CFI = .96).
60
Team Creativity
Team creativity was measured by a four-item scale adapted from Gilson and
Shalley (2004). Team leaders assessed the extent to which teams engaged in the creative
process. Sample items include “my team links ideas that originate from multiple
sources”, “my team searches for novel approaches not required at the time”, and “my
team is good at coming up with new ways of doing things”. Coefficient alpha for the
scale was 0.25 suggesting unacceptable reliability. I ran confirmatory factor analysis to
test whether the original 6-item scale fits a one-factor model or not and the model showed
a reasonable fit to the data (χ2 = 13.51, df = 9, n.s.; SRMR = .08, NNFI = .95, IFI = .97,
CFI = .97). An inter-item intercorrelation suggested the presence of error covariances
amongst the positive and negative worded items. Based on the results of confirmatory
factor analyses, I dropped two reverse scored items from the original six-item scale.
Coefficient alpha for the modified scale was 0.86. The final 4-item scale provided an
excellent fit to the data (χ2 = 2.08, df = 2, n.s.; SRMR = .03, NNFI = 1.00, IFI = 1.00,
CFI = 1.00).
61
No
Characteristics
Number
Number of Average
of teams
people
team
responding size
1
2
Country
Industry
Table 1. Sample details
USA
26
123
10.6
India
15
126
12.9
Information Technology
14
124
13.6
Aerospace
12
62
13.9
Education
7
27
5.0
Healthcare
5
29
11.8
Engineering
1
2
5.0
Retail
1
3
10.0
Finance
1
2
2.0
62
Variable
ICC (1)
ICC(2)
F-test
p-value
Rwg
(mean, SD)
TMS Specialization
0.28
0.70
3.31
0.00
0.84, 0.21
TMS Credibility
0.42
0.81
5.35
0.00
0.84, 0.30
TMS Coordination
0.10
0.40
1.67
0.01
0.55, 0.46
TMS
0.28
0.70
3.35
0.00
0.80, 0.19
TMM alternate scale
0.21
0.62
2.65
0.00
0.84, 0.20
Task Interdependence
0.14
0.50
2.00
0.00
0.74, 0.25
Team Reward
0.05
0.23
1.30
0.12
0.68, 0.34
Organizational Support
0.14
0.50
2.00
0.00
0.74, 0.21
Team viability
0.14
0.49
1.96
0.00
0.88, 0.15
Team satisfaction
0.03
0.16
1.19
0.22
0.84, 0.17
Table 2. Intraclass Coefficients (ICCs) and within-group agreement (Rwg) indices
63
Scale (items)
S-B χ2 (df)
SRMR
NNFI
IFI
CFI
TMM – alternate scale (6)
16.54(9)
0.05
0.99
0.77
0.92
Task Interdependence (5)
13.50(5)*
0.04
0.93
0.97
0.97
Organizational Support (5)
10.16(5)
0.09
0.96
0.98
0.98
Team performance (6)
14.52(6)
0.08
0.91
0.95
0.95
Team learning (5)
5.46(4)
0.12
0.90
0.97
0.96
Team viability (7)
69.46(14)*
0.07
0.95
0.97
0.97
TMS
TMS – single factor (15)
213.03(90)*
0.08
0.83
0.86
0.86
TMS – 3 first order factor
116.57(87)*
0.08
0.96
0.97
0.97
116.57(87)*
0.08
0.96
0.97
0.97
(15)
TMS – 3 first order, one
higher order factor (15)
Team creativity
Team creativity, original
13.51(9)
0.08
0.95
0.97
0.97
2.08(2)
0.03
1.00
1.00
1.00
scale, single factor (6)
Team creativity– modified
(4)
Note. S-B χ2 = Satorra-Bentler scaled chi square, df = degrees of freedom, SRMR
=Standardized root mean squared residual, NNFI = Non-normed fit index, IFI=
Incremental fit index, CFI = Comparative fit index.
*p<0.05
Table 3. Confirmatory factor analyses of the scales
64
CHAPTER 4: RESULTS
Descriptive statistics and zero-order intercorrelations for the variables are
presented in Table 4. Several significant relationships involving control variables were
observed. First, team tenure was positively related to high levels of TMS Specialization (r
= 0.32, p<0.05) and TMS credibility (r = 0.43, p<.05), and teams with longer tenure
reported a higher level of team performance (r = 0.34, p<0.05), team learning (r = 0.42,
p<0.05) and team creativity (r = 0.39, p<0.05). Second, team size was negatively related
to TMS specialization (r = -0.44, p<0.05), TMS credibility (r = -0.41, p<0.05), TMM
Average Deviation (r = -0.36, p<0.05), team performance (r = -0.33, p<0.05) and team
creativity (r = -0.49, p<0.05). Finally, country of origin affected the results. In addition,
compared to teams from India, teams from the US had higher levels of TMS
specialization (r = 0.31, p<0.05), TMS credibility (r = 0.63, p<0.05), engaged in more
interdependent tasks (r = 0.38, p<0.05), and had higher levels of team performance (r =
0.45, p<0.05) and team creativity (r=0.43, p<0.05).
Teams high on TMS specialization demonstrated high performance (r = 0.42,
p<0.05) and high team creativity (r = 0.33, p<0.05). Teams that exhibited high levels of
TMS credibility were associated with high levels of team performance (r = 0.46, p<0.05)
and team creativity (r = 0.45, p<0.05). In addition, TMS credibility was associated with
team viability (r = 0.29, p<0.05) and team creativity (r = 0.45, p<0.05). Although the
correlations were generally consistent with the proposed hypothesis, the relationships
between (1) TMS coordination and team performance and (2) TMM and team
performance, though in the expected direction, were not statistically significant.
As mentioned earlier, I operationalized TMM agreement in two different ways –
1) by computing an average deviation index and 2) by administering an equivalent 6-item
65
scale – of six major team processes. Interestingly, these two measures showed little
intercorrelation (r = -0.12, n.s.). Since both these methods were exploratory approaches, I
proceeded to test my hypothesis by examining these two methods of computing TMM
agreement. The rest of this chapter compares these two methods and describes the results.
Method 1: Team Mental Model (Average Deviation)
Relationship with Team Mental Model (TMM)
To test Hypothesis 1, I conducted a series of hierarchical regression analyses
reported in Tables 5-10. Team tenure, team size and a dummy variable representing
country were explored as control variables. Detailed results are presented in Table 5.
First, I included control variables (team tenure, team size and country) found to have
significant bivariate relationships with the dependent variables. Second, I tested whether
TMM squared, hypothesized to have direct effects (entered in step 3), had predictive
power over and above that of the TMM (entered in step 2) and the control variables
(entered in step 1). A significant change in R2 in step 3 indicates a curvilinear
relationship.
I tested all the hypotheses at a p-value of 0.10. Increasing significance level (p
values) to test the hypothesis generally increase the chance of “false positives” or Type I
error reflecting the chance that the observed relationship is not true. On the flip side, type
II error or “false negatives” reflecting the failure to detect relationships will decrease. It is
necessary to balance the Type I and Type II errors in any statistical analysis.
Hypothesis 1 predicts a ∩-shaped relationship between TMM and team
performance. This prediction was not supported (∆R2 = 0.00, n.s.). No ∩-shaped
relationship was observed between TMM and other dependent variables - team learning
66
(∆R2 = 0.00, n.s.), team creativity (∆R2 = 0.03, n.s.), team viability (∆R2 = 0.00, n.s.) or
team satisfaction (∆R2 = 0.02, n.s.). In addition, no linear relationship was observed
between TMM with team performance, team learning or team creativity. However, a
linear relationship was observed between TMM and team viability (∆R2 = 0.09, p<0.10).
A similar relationship was observed between TMM and team satisfaction (∆R2 = 0.08,
p<0.10). Relationships of TMM with team viability and team satisfaction should be
considered tenuous because of using a higher significance level (p<0.10) to test the
hypothesis.
In order to have adequate power to detect relationships between hypothesized
variables, and to limit the number of independent variables in the regression analyses, I
dropped these control variables (team tenure, team size and country of origin) from
further analysis. In further regression analyses, I included contextual variables like task
interdependence, team reward and organizational support as more theoretically
meaningful control variables. Table 6 reports the results of hierarchical regression models
by incorporating the changed control variables. Again, Hypothesis 1 predicting a ∩shaped relationship between TMM and team performance was not supported (∆R2 = 0.00,
n.s.). No ∩-shaped relationship was observed between TMM and other dependent
variables - team learning (∆R2 = 0.00, n.s.), team creativity (∆R2 = 0.00, n.s.), team
viability (∆R2 = 0.01, n.s.) or team satisfaction (∆R2 = 0.01, n.s.). In addition, no linear
relationship was observed for TMM with team performance, team learning, team
creativity, team viability or team satisfaction. In summary, this set of results suggests lack
of support for H1, predicting a curvilinear relationship between TMM and team
performance.
67
Relationship with TMS
Tables 7-10 report a series of hierarchical regression models used to test
Hypothesis 2. First, I included the control variables (task interdependence, team reward
and organizational support) found to have significant bivariate relationships with the
dependent variables. Second, I tested whether TMS or its components (entered in step 3)
had predictive power over TMM (entered in step 2) and the control variables (entered in
step 1). A significant change in R2 in step 3 indicates a main effect. Hypothesis 2a-c
predicts that TMS components (specialization, coordination and credibility) will be
positively related to team performance. Before testing for TMS subcomponents, I
examined the TMS construct first and its impact on team performance. Based on results
from Step 3 of the model in Table 7, TMS had a weak positive relationship with team
performance (∆R2 = 0.06, p<0.10). There was an absence of relationships between TMS
and team learning, team creativity, team viability and team satisfaction. Detailed results
are presented in Table 7.
Next, I examined H2a that predicts that TMS specialization will be positively
related to team performance. I tested whether TMS specialization (entered in step 3) had
predictive power over TMM (entered in step 2) and control variables (entered in step 1).
A significant change in R2 in step 3 indicates a main effect. This prediction was not
supported (∆R2 = 0.06, n.s.). Further, based on results from step 3 of Table 8, no
relationship was observed for TMS specialization with team learning (∆R2 = 0.01, n.s.),
team creativity (∆R2 = 0.02, n.s.), team viability (∆R2 = 0.00, n.s.) or team satisfaction
(∆R2 = 0.02, n.s.). Detailed results are presented in Table 8.
68
Next, I investigated H2b that predicts TMS coordination will be positively related
to team performance. I tested whether TMS coordination (entered in step 3) had
predictive power over TMM (entered in step 2) and the control variables (entered in step
1). A significant change in R2 in step 3 indicates a main effect. This prediction was not
supported (∆R2 = 0.00, n.s.). Further, based on results from step 3, no relationship was
observed for TMS coordination with team learning (∆R2 = 0.04, n.s.) or team creativity
(∆R2 = 0.02, n.s.). A linear relationship was observed for TMS coordination with team
viability (∆R2 = 0.05, p<0.05) and team satisfaction (∆R2 = 0.04, p<0.10). Detailed
results are reported in Table 9.
Next, I investigated H2c that predicts TMS credibility will be positively related to
team performance. I tested whether TMS credibility (entered in step 3) had predictive
power over TMM (entered in step 2) and the control variables (entered in step 1). A
significant change in R2 in step 3 indicates a main effect. This prediction was supported
(∆R2 = 0.07, p<0.10). TMS credibility also predicted team creativity (∆R2 = 0.07,
p<0.10). Further, based on results from step 3, no relationship was observed for TMS
credibility with team learning (∆R2 = 0.05, n.s.), team viability (∆R2 = 0.00, n.s.) or team
satisfaction (∆R2 = 0.02, n.s.). Detailed results are reported in Table 10. Overall, these
sets of results provide partial support for Hypothesis 2c, but fail to support Hypothesis 2a
and 2b.
Relationship between TMM and TMS
Table 7 also reports a series of hierarchical regression models used to test
Hypothesis 3. First, I included the control variables (task interdependence, team reward,
organizational support) found to have significant bivariate relationships with the
69
dependent variables. Second, I tested whether the cross-product interaction term had
direct effects (entered in step 4) and predictive power over and above that of TMS
(entered in step 3), TMM (entered in step 2) and the control variables (entered in step 1).
Third, I centered independent variables (TMM & TMS) to reduce multicollinearity
during testing for interactions (Aiken & West, 1991). A significant change in R2 in step 4
indicates a moderation effect. Hypothesis 3 predicts that TMM moderates the relationship
between TMS and team performance such that the relationship is negative when TMM is
low but positive when TMM is high. This prediction was not supported (∆R2 = 0.03,
n.s.).
Further, no moderation effect of TMM was detected for the TMS related to team
learning (∆R2 = 0.05, n.s.), team creativity (∆R2 = 0.08, p<0.10), team viability (∆R2 =
0.01, n.s.) or team satisfaction (∆R2 = 0.00, n.s.). However, TMM moderated the
relationship between TMS and team creativity (∆R2 = 0.08, p<0.10). Contrary to
expectations, TMM aided team creativity under conditions of low TMS, but hampered
team creativity under high TMS conditions. Figure 3 graphically depicts this relationship.
Team creativity was highest under conditions of high TMS and low TMM, and worst
under conditions of low TMS and low TMM.
Post-hoc analyses: TMS Components
Although the previous set of results failed to support the moderating role of TMM
on TMS and team performance, I explored the possibility that TMM might interact with
TMS components (TMS specialization, TMS credibility and TMS coordination) and team
effectiveness dimensions. This was indirectly supported by the initial evidence that a
three-factor confirmatory factor analyses fit the data better than a single factor TMS, the
70
presence of three distinct components (Table 3). More support for an exploratory
approach came from the evidence that TMS components behaved differently with team
performance as reported earlier in the bivariate correlation analyses (Table 4). Thus, I
engaged in exploration of how these distinct components behave in relation to my
proposed hypothesis. Engaging in post-hoc tests raises the possibility of alpha-inflation.
In other words, the more tests I conduct at a given significance level (say α= .05), the
more likely the chance to claim a significant result when there isn’t one (i.e., a Type I
error increasingly exceeds the nominal level). Hence, care should be taken in interpreting
these sets of results.
Similar to Hypothesis 1 that predicted curvilinear effect with TMM and team
performance, I wondered about the possibility of a curvilinear relationship between TMS
components and the dependent variables. To test these hypotheses, I ran a series of
hierarchical regression analyses. This involved a three step process. First, I included the
control variables (task interdependence, team reward and organizational support).
Second, I entered TMS specialization. Third, I tested whether TMS specialization
squared (entered in step 3) had predictive power over and above that of TMS
specialization (entered in step 2) and control variables (entered in step 1). In addition, I
centered the TMS specialization scale as suggested by Aiken and West (1991). A
significant change in R2 in step 3 indicates a curvilinear relationship. Detailed results
from Table 11 suggests that a curvilinear relationship exists between TMS specialization
and team creativity (∆R2 = 0.07, p<0.10). No such relationship was observed between
TMS specialization and the other dependent variables – team performance, team learning,
team viability and team satisfaction. Figure 4 graphically depicts the curvilinear
71
relationship between TMS specialization and team creativity. The effect was such that it
takes a fairly high level of TMS specialization before teams experience increased team
creativity. Even an average level of TMS specialization doesn’t help team creativity. The
level of TMS has to exceed a certain threshold level before it impacts creativity.
Next, I ran a similar series of hierarchical regression to test the relationship
between TMS coordination and the dependent variables. In particular, I tested whether
TMS coordination squared (entered in step 3) had predictive power over and above TMS
coordination (entered in step 2) and the control variables (entered in step 1). I also
centered the TMS coordination scale to minimize multicollinearity (Aiken & West,
1991). Detailed results are presented in Table 12. Curvilinear relationships were observed
for TMS coordination with team viability (∆R2 = 0.09, p<0.01) and team satisfaction
(∆R2 = 0.04, p<0.10). A high level of TMS coordination aided team viability. Figure 5
graphically depicts this relationship. The effect was such that it takes a fairly high level of
TMS coordination before teams experience increased team viability. Even an average
level of TMS coordination doesn’t help team viability and the level of TMS coordination
has to exceed a certain threshold level before it impacts viability. A similar pattern was
observed between TMS coordination and team satisfaction. Team satisfaction increased
beyond a certain threshold level of TMS coordination. Figure 6 graphically depicts this
relationship.
Next, I ran a similar series of hierarchical regressions to test the relationship
between TMS credibility and the dependent variables. In particular, I tested whether TMS
credibility squared (entered in step 3) had predictive power over and above that of TMS
credibility (entered in step 2) and the control variables (entered in step 1). I also centered
72
the TMS credibility scale to minimize multicollinearity (Aiken & West, 1991). Detailed
results are presented in Table 13. A nonlinear relationship was observed between TMS
credibility and team performance (∆R2 = 0.08, p<0.05). No curvilinear relationship was
observed with other dependent variables. The non-linear effect was such that the
increasing level of TMS credibility helps increase team performance. However, once
TMS credibility goes past average level, there is no additional benefit in terms of team
performance. Stated another way, teams perform badly under conditions of low TMS
credibility. Team performance is not a problem under average or high levels of TMS
credibility. Figure 7 graphically depicts this relationship.
Moderating impact of TMS Components
I explored the possibility of the moderating impact of TMM on the relationship
between TMS components and the dependent variables. I investigated this possibility by
running a series of hierarchical regression analyses as reported in Tables 8-10.
Specifically, I entered the relevant interaction term (e.g. TMS Specialization X TMM) in
step 4 of the regression equation of Table 8. No moderation effect was observed for
TMM on TMS specialization and team performance. However, an interaction effect was
observed between TMM and TMS specialization on team creativity (∆R2 = 0.08, p<.10).
Figure 8 graphically depicts this relationship.
According to figure 8, TMM generally resulted in higher levels of team creativity.
As expected, team creativity was lowest when both TMM and TMS Specialization were
low. The interaction was such that when TMS was low, TMM mattered. When TMS
specialization was high, there was hardly any difference in team creativity. When TMS
73
specialization was low, high TMM facilitated team creativity while low TMM lowered
team creativity.
Similar analyses reported in figure 9 reveals an interaction between TMM and
TMS coordination on team performance (∆R2 = 0.12, p<.05). As expected, teams high on
TMM demonstrated better than average performance under conditions of high TMS
coordination. Team performance was equally good under conditions of low TMM
coupled with low TMS coordination. This is puzzling and difficult to explain. However,
team performance went down under conditions of low TMM and high TMS coordination.
Again contrary to expectations, teams high on TMM demonstrated a lower than average
team performance when coupled with high levels of TMS coordination, while teams low
on TMM demonstrated a better than average performance when coupled with low levels
of TMS coordination. TMS coordination and TMM thus demonstrated a compensatory
relationship when predicting team performance.
Similar analyses reported in Table 10 reveal an interaction between TMM and
TMS credibility with team performance (∆R2 = 0.14, p<.01), team learning (∆R2 = 0.20,
p<.01), and team creativity (∆R2 = 0.14, p<.01). Figures 10-12 depict these relationships.
Figure 10 illustrates that teams high on TMS credibility generally demonstrated
better team performance. Teams had their worst performance when they were low on
both TMM and TMS credibility. However, teams demonstrated superior team
performance under conditions of high levels of TMS credibility coupled with low levels
of TMM. This result was counter-intuitive. Improvements in TMS credibility aided teams
lowest on TMM levels. Overall, the interaction pattern was evident under conditions of
low TMM. Teams high on TMM delivered average performance.
74
Figure 11 illustrates that teams high on TMS credibility engaged in more learning
behaviors than the teams low on TMS credibility. Teams engaged in the least learning
when they were low on both TMM and TMS credibility. However, teams engaged in
more learning behaviors under conditions of high levels of TMS credibility coupled with
low levels of TMM. This result was counter-intuitive. Improvements in TMS credibility
aided teams lowest on TMM levels. Overall, the interaction pattern was evident under
conditions of low TMM. Teams high on TMM engaged in an average level of team
learning.
Figure 12 illustrates that teams high on TMS credibility engaged in more
creativity than those teams low on TMS credibility. Teams were least creative when they
were low on both TMM and TMS credibility. However, teams engaged in behaviors that
are more creative when they experienced low levels of TMM and high levels of TMS
credibility. This result was counter-intuitive. In other words, the interaction pattern was
evident under conditions of low TMM. Teams high on TMM engaged in an average level
of team learning. Overall, these sets of results indicate that TMS has its largest impact in
predicting team performance, teams learning and team creativity on teams with low
TMM.
Method 2: Team Mental Model (Alternate Scale)
Since this was an exploratory study of TMM in a field setting involving teams
from multiple industries, I operationalized the TMM scale in two different ways – one
using the average deviation of 6 rank-ordered team processes and the other a 6-item
Likert type equivalent scale. I ran the entire analysis as described in the previous section
75
using the alternate Likert scale developed for this study. The following section details the
results.
Relationship with Team Mental Model (TMM)
To test Hypothesis 1, I conducted a series of hierarchical regression analyses as
reported in Table 14. Team tenure, team size and a dummy variable representing country
were explored as control variables. First, I included control variables (team tenure, team
size and country) found to have significant bivariate relationships with the dependent
variables. Second, I tested whether TMM squared hypothesized to have direct effects
(entered in step 3) had predictive power over and above that of TMM (entered in step 2)
and the control variables (entered in step 1). Third, I centered the TMM scale to reduce
multicollinearity when testing for interactions (Aiken & West, 1991). A significant
change in R2 in step 3 indicates a curvilinear relationship.
Hypothesis 1 predicted a ∩-shaped relationship between TMM and team
performance. This prediction was not supported (∆R2 = 0.01, n.s.). No ∩-shaped
relationship was observed between TMM and the other dependent variables - team
learning (∆R2 = 0.00, n.s.), team creativity (∆R2 = 0.01, n.s.), team viability (∆R2 = 0.01,
n.s.) and team satisfaction (∆R2 = 0.00, n.s.). However, there were linear relationships for
TMM with team performance (∆R2 = 0.06, p<0.10), team creativity (∆R2 = 0.08, p<0.05),
team viability (∆R2 = 0.42, p<0.01) and team satisfaction (∆R2 = 0.40, p<0.01). In
addition, control variables (team tenure, team size and country of origin) as a block
explained a sizeable proportion of variance in team performance (∆R2 = 0.32, p<0.01),
team learning (∆R2 = 0.27, p<0.01) and team creativity (∆R2 = 0.38, p<0.01).
Specifically, team size was negatively related to team performance (β = -.27, p<0.10) and
76
team creativity (β = -.27, p<0.10). Further, teams from the US were rated higher on team
performance (β = .37, p<0.10) than those from India.
In order to have adequate power to detect relationships between hypothesized
variables, and to limit the number of independent variables in the regression analyses, I
dropped these control variables (team tenure, team size and country of origin) from
further analysis. Dropping these control variables did not impact the substantial
interpretation of results. In further regression analysis, I included contextual variables
like task interdependence, team reward and organizational support as more theoretically
meaningful control variables. Table 15 reports the results of hierarchical regression
models by incorporating the changed control variables. Again, Hypothesis 1, predicting a
∩-shaped relationship between TMM and team performance, was not supported (∆R2 =
0.01, n.s.). No ∩-shaped relationship was observed between TMM and other dependent
variables - team learning (∆R2 = 0.00, n.s.), team creativity (∆R2 = 0.02, n.s.), team
viability (∆R2 = 0.01, n.s.) and team satisfaction (∆R2 = 0.01, n.s.). However, there was a
linear relationship for TMM with team learning (∆R2 = 0.11, p<0.05), team viability (∆R2
= 0.07, p<0.05) and team satisfaction (∆R2 = 0.08, p<0.05). In addition, no linear
relationship was observed for TMM with team performance or team creativity. In
summary, this set of results suggests lack of support for H1 which predicted a curvilinear
relationship between TMM and team performance.
Relationship with TMS
Tables 16-19 report a series of hierarchical regression models used to test
Hypothesis 2. First, I included the control variables (task interdependence, team reward
and organizational support) found to have significant bivariate relationships with the
77
dependent variables. Second, I tested whether TMS or its components (entered in step 3)
had predictive power over TMM (entered in step 2) and the control variables (entered in
step 1). A significant change in R2 in step 3 indicates a main effect. Hypotheses 2a-c
predict that TMS components (specialization, coordination and credibility) will be
positively related to team performance. Before testing for TMS subcomponents, I first
examined the TMS construct and its impact on team performance. Based on results from
Step 3 of the model in Table 16, TMS had weak positive relationships with team
performance (∆R2 = 0.06, p<0.10) and team creativity (∆R2 = 0.06, p<0.10). There was
no relationship for TMS with team learning, team viability or team satisfaction. Detailed
results are presented in Table 16.
Next, I examined H2a that predicted TMS specialization to be positively related
to team performance. I tested whether TMS specialization (entered in step 3) had
predictive power over TMM (entered in step 2) and the control variables (entered in step
1). A significant change in R2 in step 3 indicates a main effect. This prediction was not
supported (∆R2 = 0.06, n.s.). Further, based on results from step 3 of Table 17, no
relationship was observed for TMS specialization with team learning (∆R2 = 0.01, n.s.),
team creativity (∆R2 = 0.02, n.s.), team viability (∆R2 = 0.00, n.s.) or team satisfaction
(∆R2 = 0.00, n.s.). Detailed results are presented in Table 17.
Next, I investigated H2b that predicted TMS coordination to be positively related
to team performance. I tested whether TMS coordination (entered in step 3) had
predictive power over TMM (entered in step 2) and the control variables (entered in step
1). A significant change in R2 in step 3 indicates a main effect. This prediction was not
supported (∆R2 = 0.00, n.s.). Further, based on results from step 3, no relationship was
78
observed for TMS coordination with team learning (∆R2 = 0.01, n.s.), team creativity
(∆R2 = 0.03, n.s.), team viability (∆R2 = 0.01, n.s.) or team satisfaction (∆R2 = 0.00, n.s.).
Detailed results are reported in Table 18.
Next, I investigated H2c that predicted TMS credibility to be positively related to
team performance. I tested whether TMS credibility (entered in step 3) had predictive
power over TMM (entered in step 2) and the control variables (entered in step 1). A
significant change in R2 in step 3 indicates a main effect. This prediction was supported
(∆R2 = 0.05, p<0.05). TMS credibility also predicts team creativity (∆R2 = 0.05, p<0.10).
Further, based on results from step 3, no relationships were observed for TMS credibility
with team learning (∆R2 = 0.02, n.s.), team viability (∆R2 = 0.01, n.s.) or team
satisfaction (∆R2 = 0.00, n.s.). Detailed results are reported in Table 19. Overall, these
sets of results partially support Hypothesis 2c, but fail to support Hypotheses 2a and 2b.
Relationship between TMM and TMS
Table 16 reports a series of hierarchical regression models used to test Hypothesis
3. First, I included the control variables (task interdependence, team reward,
organizational support) found to have significant bivariate relationships with the
dependent variables. Second, I tested whether the cross-product interaction term had
direct effects (entered in step 4) and predictive power over and above that of TMS
(entered in step 3), TMM (entered in step 2) and the control variables (entered in step 1).
Third, I centered independent variables (TMM & TMS) to reduce multicollinearity
during testing for interactions (Aiken & West, 1991). A significant change in R2 in step 4
indicates a moderation effect. Hypothesis 3 predicts that TMM moderates the relationship
between TMS and team performance such that the relationship is negative when TMM is
79
low but positive when TMM is high. This prediction was not supported (∆R2 = 0.01,
n.s.).
Further, no moderation effect of TMM was detected for the relationships of TMS
with team learning (∆R2 = 0.00, n.s.), team viability (∆R2 = 0.01, n.s.) or team
satisfaction (∆R2 = 0.01, n.s.). However, TMM moderated the relationship between TMS
and team creativity (∆R2 = 0.09, p<0.05). TMS generally aided team creativity
irrespective of level of TMM. When TMS was low, team creativity was better under
conditions of low TMM than under high TMM. Team creativity was worst under
conditions of low TMS and high TMM. It appeared that TMM seemed to hamper team
creativity especially under conditions of low TMS. Figure 13 graphically depicts this
relationship.
Post-hoc analyses
Although the previous set of results failed to support the moderating role of TMM
on the relationship between TMS and team performance, I explored the possibility that
TMM might interact differently with TMS components (TMS Specialization, TMS
Credibility and TMS Coordination) and team effectiveness dimensions. This possibility
was supported by the initial evidence that a three-factor confirmatory factor analyses fit
the data better than a single factor TMS with the presence of three distinct components
(Table 2). More support for this exploratory approach came from the evidence that TMS
components behaved differently with team performance as reported earlier in the
bivariate correlation analyses (Table 3). Thus, I engaged in exploration of how these
distinct components behave in relation to my proposed hypothesis.
80
Moderating impact of TMS components
I explored the possibility of the moderating impact of TMM on the relationship
between TMS components and the dependent variables. I investigated this possibility by
running a series of hierarchical regression analyses reported in Tables 17-19.
Specifically, I entered the relevant interaction term (e.g., TMS Specialization X TMM) in
step 4 of the regression equation of Table 17. TMM did not moderate the relationship
between TMS specialization and team performance. However, an interaction effect was
observed between TMM and TMS specialization on team creativity (∆R2 = 0.12, p<.05).
Figure 14 graphically depicts this relationship.
According to Figure 14, higher levels of TMS specialization resulted in higher
levels of team creativity. Contrary to expectations, team creativity was lowest under
conditions of high TMM. Moreover, teams demonstrated their lowest level of team
creativity when TMM was high but TMS specialization was low. This result was
puzzling in light of Hypothesis 3. It appears that high TMM has a detrimental impact on
team creativity.
Similar analyses reported in Table 18 reveal an interaction between TMM and
TMS coordination on team viability (∆R2 = 0.05, p<.05). Figure 15 shows the
relationship graphically. As expected, team viability increased under conditions of high
TMM and high TMS coordination. Conversely, team viability was lowest under
conditions of low TMM coupled with high levels of TMS coordination. High levels of
TMM increased team viability, or their intention to maintain themselves. TMS
coordination helped increase team viability under conditions of high TMM.
81
Similar analyses reported in Table 19 reveal an interaction between TMM and
TMS credibility for team performance (∆R2 = 0.14, p<.01), team learning (∆R2 = 0.20,
p<.01), and team creativity (∆R2 = 0.14, p<.01). Figure 16 depicts the relationship for
creativity. The figure reveals that team creativity remains at a fairly high level under
conditions of low TMM. Team creativity decreases when teams demonstrate low TMS
credibility coupled with high TMM. TMS credibility generally helps improve team
creativity except under conditions of low TMM. At high levels of TMS credibility,
irrespective of TMM level, team creativity is quite high.
Table 20 details the key findings of all analyses, including similarities and
differences, from analyses obtained by following two different methods of TMM
operationalization – the average deviation method and the 6-item Likert scale. Support
for the results has been marked by figures depicted previously in the manuscript.
Variable
Mean
SD
1
2
3
4
5
6
7
8
1. Team tenure
2.91
2.16
-
2. Countrya
0.68
0.47
0.53
10.90
7.42
-0.16
-0.24
-
4. TMS specialization
4.16
0.47
0.32
0.31
-0.44
(.61)
5. TMS credibility
4.00
0.54
0.43
0.63
-0.41
0.68
(.77)
6. TMS coordination
3.91
0.43
0.17
-0.16
-0.12
0.06
0.20
(.66)
7. TMS
4.02
0.36
0.43
0.39
-0.45
0.80
0.88
0.53
(.79)
-1.04
0.29
-0.12
0.22
-0.36
0.21
0.14
-0.15
0.11
-
3.88
0.52
-0.17
-0.50
0.09
0.17
0.00
0.64
0.33
-0.12
9
Control variables
3. Team size
-
Independent variables
8. TMM (AD)
9. TMM alt scale
(.82)
Note. a Country (0=India, 1=US), N=41 teams. Scale reliabilities are on the diagonal in parentheses. The 95% confidence interval for
correlations greater than or equal to 0.29 does not include 0 (.01<0.29<.57).b Reported by team managers, other measures obtained
from team members.
82
Table 4. Means, Standard Deviations and Inter-correlations among study variables
Variable
Mean
SD
1
2
3
4
5
6
7
8
9
10. Task interdependence
3.45
0.61
0.32
0.38
-0.08
0.29
0.37
0.16
0.38
-0.05
0.04
11. Team reward
3.86
0.46
0.04
0.26
-0.21
0.51
0.54
0.25
0.59
0.10
0.38
12. Org support
3.52
0.53
0.15
-0.25
-0.06
0.20
0.24
0.63
0.45
-0.05
0.64
13. Team viability
3.98
0.56
-0.07
-0.14
-0.04
0.23
0.29
0.59
0.48
0.19
0.67
14. Team satisfaction
4.21
0.43
0.11
-0.09
-0.01
0.19
0.22
0.57
0.42
0.13
0.64
15. Team performanceb
4.08
0.51
0.34
0.45
-0.33
0.42
0.46
0.17
0.48
0.06
0.09
16. Team learningb
3.95
0.72
0.42
0.43
-0.24
0.16
0.25
-0.05
0.18
0.11
-0.12
17. Team creativityb
4.03
0.78
0.39
0.43
-0.49
0.33
0.45
0.24
0.47
0.13
0.05
Contextual variables
Dependent variables
Note. a Country (0=India, 1=US), N=41 teams. Scale reliabilities are on the diagonal in parentheses. The 95% confidence interval for
correlations greater than or equal to 0.29 does not include 0 (.01<0.29<.57).b Reported by team managers, other measures obtained
from team members.
Table 4. Continued
83
Variable
10
11
12
13
14
15
16
17
Contextual variables
10. Task interdependence
(.74)
11. Team reward
0.44
(.58)
12. Org support
0.02
0.29
(.70)
13. Team viability
0.12
0.46
0.66
(.86)
14. Team satisfaction
0.32
0.04
0.63
0.76
(.73)
15. Team performanceb
0.24
0.39
0.27
0.10
0.12
(.83)
16. Team learningb
0.00
0.06
0.20
-0.06
0.00
0.73
(.63)
17. Team creativityb
0.24
0.36
0.26
0.14
0.17
0.77
0.70
Dependent variables
(.86)
Note. a Country (0=India, 1=US), N=41 teams. Scale reliabilities are on the diagonal in
parentheses. The 95% confidence interval for correlations greater than or equal to 0.29 does
not include 0 (.01<0.29<.57).b Reported by team managers, other measures obtained from team
members.
Table 4. Continued
84
Team performance
Predictor
Team tenure
Team size
Country
Model 2
Model 3
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
.11
.09
.09
.29†
.31†
.31†
.21
.20
.22
-.27†
-.29†
-.29†
-.13
-.12
-.11
-.41**
-.42**
-.39*
.37*
.39*
.39*
.26
.24
.24
.23
.24
.25
.05
.04
-.04
-.15
-.07
TMM squared
∆R2
Team creativity
Model 1
TMM
R2
Team learning
-.06
-.01
.32**
.32**
.32**
.00
.00
.03
.27*
.27*
.27*
.00
.00
.19
.38**
.38**
.41**
.00
.03
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 5. Results of Hierarchical Regression Analyses testing Hypothesis 1 with team tenure, team size and country of origin as control
variables: TMM Average Deviation
85
Team viability
Predictor
Team satisfaction
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Team tenure
-.04
.07
.06
.17
.27
.25
Team size
-.12
-.01
-.02
-.07
.03
.01
Country
-.12
-.22
-.22
-.18
-.27
-.28
.34†
.39†
TMM
TMM squared
R2
∆R2
.32†
-.08
.03
.12
.12
.09†
.00
.42†
-.17
.03
.11
.13
.08†
.02
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 5. Continued
86
Team performance
Predictor
Team learning
Team creativity
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Task interdependence
.11
.12
.11
.00
.02
.02
.12
.14
.15
Team reward
.28
.27
.30
-.01
-.03
-.02
.25
.23
.21
Organizational support
.19
.19
.19
.21
.22
.22
.18
.19
.20
.04
.12
.13
.16
.13
.07
TMM
TMM squared
R2
∆R2
-.13
.19
.19
.20
.00
.01
-.05
.04
.06
.06
.02
.00
.09
.16†
.18†
.18†
.02
.00
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 6. Results of Hierarchical Regression Analyses testing Hypothesis 1 with task interdependence, team reward and organizational
support as control variables: TMM Average Deviation
87
Team viability
Predictor
Team satisfaction
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Task interdependence
-.03
-.01
-.02
.26*
.29*
.27*
Team reward
.31*
.27*
.30*
.10
.07
.12
Organizational support
.57**
.59**
.58**
.59**
.61**
.60**
.19
.27†
.17
.30*
TMM
TMM squared
R2
∆R2
-.14
.51**
.55**
.56**
.04
.01
-.22
.50**
.52**
.55**
.02
.03
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 6. Continued
88
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11
.12
.05
.04
.00
.02
-.01
-.02
.12
.14
.07
.06
Team reward
.28
.27
.14
.15
-.01
-.03
-.09
-.07
.25
.23
.09
.11
Organizational support
.19
.19
.08
.11
.21
.22
.17
.22
.18
.19
.08
.13
.04
.01
.02
.13
.11
.13
.13
.10
.12
.34†
.37†
.15
.18
.34
.39†
TMM
TMS
TMM X TMS
R2
∆R2
-.16
.19*
.19*
.25*
.28*
.00
.06†
.03
-.29†
-.25
.04
.06
.07
.12
.02
.01
.05
.16†
.18†
.24†
.32*
.02
.06
.08†
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 7. Results of Hierarchical Regression Analyses examining impact of TMS on team outcomes: TMM Average Deviation
89
Team viability
Predictor
Team satisfaction
Model Model Model Model Model Model Model Model
1
2
3
4
Task interdependence
-.03
-.01
-.02
Team reward
.31*
.27*
Organizational support
.57**
TMM
1
2
3
4
-.01
.26*
.29*
.30*
.30*
.25
.24
.10
.07
.09
.09
.59**
.57**
.54**
.59**
.61**
.63**
.63**
.19
.18
.17
.17
.17
.17
.07
.05
-.05
-.05
TMS
TMM X TMS
R2
∆R2
.12
.51**
.55**
.55**
.56**
.04
.00
.01
-.02
.50**
.52**
.53**
.53**
.02
.01
.00
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 7. Continued
90
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11
.12
.09
.07
.00
.02
.00
-.01
.12
.14
.12
.10
Team reward
.28
.27
.15
.25
-.01
-.03
-.09
.00
.25
.23
.16
.28
Organizational support
.19
.19
.17
.18
.21
.22
.21
.22
.18
.19
.18
.20
.04
-.01
.06
.13
.10
.16
.13
.10
.18
.29
.31†
.15
.17
.16
.19
TMM
TMS spec
TMM X TMS spec
R2
∆R2
-.26
.19*
.19*
.25*
.30*
.00
.06
.05
-.33†
-.24
.04
.06
.07
.11
.02
.01
.04
.16†
.18†
.20†
.28*
.02
.02
.08†
Note. N=41 teams, Regression coefficients are standardized betas, spec = specialization.
**p<0.01, *p<0.05, †p<0.10
Table 8. Results of Hierarchical Regression Analyses testing Hypothesis 2a and 3: TMM Average Deviation
91
Team viability
Predictor
Team satisfaction
Model Model Model Model Model Model Model Model
1
2
3
4
Task interdependence
-.03
-.01
.01
Team reward
.31*
.27*
Organizational support
.57**
TMM
1
2
3
4
.02
.26*
.29*
.30*
.30*
.31*
.25
.10
.07
.12
.12
.59**
.60**
.59**
.59**
.61**
.62**
.62**
.19
.21†
.16
.17
.19
.19
-.09
-.11
-.13
-.13
TMS specialization
TMM X TMS spec
R2
∆R2
.17
.51**
.55**
.55**
.57**
.04
.00
.02
.00
.50**
.52**
.54**
.54**
.02
.02
.00
Note. N=41 teams, Regression coefficients are standardized betas, spec = specialization.
**p<0.01, *p<0.05, †p<0.10
Table 8. Continued
92
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11
.12
.13
.09
.00
.02
.06
.03
.12
.14
.13
.12
Team reward
.28
.27
.28
.38*
-.01
-.03
-.02
.04
.25
.23
.22
.26
Organizational support
.19
.19
.23
.17
.21
.22
.39†
.35
.18
.19
.14
.12
.04
.04
.09
.13
.09
.13
.13
.14
.15
-.06
-.15
-.28
-.33
.10
.07
TMM
TMS coord
TMM X TMS coord
R2
∆R2
.38*
.19*
.19*
.19*
.31*
.00
.00
.12*
.25
.04
.06
.10
.15
.02
.04
.05
.12
.16†
.18†
.18†
.20†
.02
.02
.02
Note. N=41 teams, Regression coefficients are standardized betas, coord = coordination.
**p<0.01, *p<0.05, †p<0.10
Table 9. Results of Hierarchical Regression Analyses testing Hypothesis 2b and 3: TMM Average Deviation
93
Team viability
Predictor
Team satisfaction
Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
Task interdependence
-.03
-.01
-.05
-.04
.26*
.29*
.25†
.25†
Team reward
.31*
.27*
.26*
.25†
.10
.07
.06
.06
Organizational support
.57**
.59**
.39**
.40**
.59**
.61**
.46**
.46**
.19
.23*
.22†
.17
.20†
.19
.32*
.34*
.25†
.25
TMM
TMS coord
TMM X TMS coord
R2
∆R2
-.07
.51**
.55**
.60**
.61**
.04
.05*
.01
-.01
.50**
.52**
.56**
.56**
.02
.04†
.00
Note. N=41 teams, Regression coefficients are standardized betas, coord = coordination.
**p<0.01, *p<0.05, †p<0.10
Table 9. Continued
94
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11
.12
.06
.02
.00
.02
-.04
-.09
.12
.14
.08
.03
Team reward
.28
.27
.14
.14
-.01
-.03
-.15
-.16
.25
.23
.09
.09
Organizational support
.19
.19
.15
.21
.21
.22
.19
.25
.18
.19
.16
.21
.04
.01
.01
.13
.09
.09
.13
.09
.09
.32†
.41*
.29
.39*
.33†
.41*
TMM
TMS cred
TMM X TMS cred
R2
∆R2
-.39**
.19*
.19*
.26*
.40**
.00
.07†
.14**
-.46**
.04
.06
.11
.31*
.02
.05
.20**
-.39**
.16†
.18†
.25*
.39**
.02
.07†
.14**
Note. N=41 teams, Regression coefficients are standardized betas, cred = credibility.
**p<0.01, *p<0.05, †p<0.10
Table 10. Results of Hierarchical Regression Analyses testing Hypothesis 2c and 3: TMM Average Deviation
95
Team viability
Predictor
Team satisfaction
Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
Task interdependence
-.03
-.01
.00
.02
.26*
.29*
.31*
.31*
Team reward
.31*
.27*
.28†
.29†
.10
.07
.13
.13
Organizational support
.57**
.59**
.59**
.57**
.59**
.61**
.63**
.63**
.19
.19
.19
.17
.18
.18
-.03
-.06
-.14
-.13
TMM
TMS cred
TMM X TMS cred
R2
∆R2
.13
.51**
.55**
.55**
.56**
.04
.00
.01
-.04
.50**
.52**
.54**
.54**
.02
.02
.00
Note. N=41 teams, Regression coefficients are standardized betas, cred = credibility.
**p<0.01, *p<0.05, †p<0.10
Table 10. Continued
96
Team performance
Predictor
Team learning
Team creativity
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Task interdependence
.11
.09
.09
.00
-.01
.00
.12
.11
.13
Team reward
.28
.15
.15
-.01
-.08
-.05
.25
.17
.23
Organizational support
.19
.17
.17
.21
.20
.23
.18
.17
.23
.29†
.29
.17
.22
.18
.28
TMS spec
TMS spec squared
R2
∆R2
.00
.19*
.25*
.25*
.06†
.00
.32†
.18
.04
.06
.09
.02
.03
.16†
.19†
.26*
.03
.07†
Note. N=41 teams, Regression coefficients are standardized betas, spec = specialization.
**p<0.01, *p<0.05, †p<0.10
Table 11. Post-hoc Hierarchical Regression Analyses: TMS specialization and team outcomes
97
Team viability
Predictor
Task interdependence
Model 1
-.03
Model 2
-.03
Team satisfaction
Model 3
Model 1
Model 2
Model 3
-.03
.26*
.27*
.27†
Team reward
.31*
.33*
.31
.10
.14
.12
Organizational support
.57**
.57**
.56**
.59**
.60**
.58**
TMS spec
-.05
TMS spec squared
R2
∆R2
-.07
-.08
-.07
.51**
.51**
.52**
.00
.01
-.10
-.08
.50**
.50**
.51**
.00
.01
Note. N=41 teams, Regression coefficients are standardized betas, spec = specialization.
**p<0.01, *p<0.05, †p<0.10
Table 11. Continued
98
Team performance
Predictor
Team learning
Team Creativity
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Task interdependence
.11
.12
.11
.00
.05
.04
.12
.11
.10
Team reward
.28
.28
.24
-.01
-.01
-.02
.25
.25
.22
Organizational support
.19
.23
.27
.21
.18
.14
.16
-.07
-.10
.07
.05
TMS coord
TMS coord squared
R2
∆R2
.39†
-.29
.10
.19*
.19*
.20*
.00
.01
.41†
-.31
.05
.04
.09
.09
.05
.00
.07
.16†
.17†
.17†
.01
.00
Note. N=41 teams, Regression coefficients are standardized betas, coord = coordination.
**p<0.01, *p<0.05, †p<0.10
Table 12. Post-hoc Hierarchical Regression Analyses: TMS coordination and team outcomes
99
Team viability
Predictor
Task interdependence
Team satisfaction
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
-.03
-.07
-.12
.26*
.23†
.20
Team reward
.31*
.31*
.17
.10
.10
.01
Organizational support
.57**
.39*
.53**
.59**
.46**
.55**
.28†
.18
.22
.15
TMS coord
TMS coord squared
R2
∆R2
.23†
.35**
.51**
.56**
.65**
.05†
.09**
.50**
.52**
.56**
.02
.04†
Note. N=41 teams, Regression coefficients are standardized betas, coord = coordination.
**p<0.01, *p<0.05, †p<0.10
Table 12. Continued
100
Team performance
Predictor
Team learning
Team creativity
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Task interdependence
.11
.06
.10
.00
-.05
-.01
.12
.06
.07
Team reward
.28
.14
.05
-.01
-.14
-.22
.25
.10
.09
Organizational support
.19
.15
.24
.21
.17
.25
.18
.14
.16
.32†
.25
.30
.24
.34†
.33†
TMS cred
TMS cred squared
R2
∆R2
-.32*
.19*
.26*
.34*
.07†
.08*
-.28
.04
.10
.17
.06
.07
-.04
.16†
.24*
.24*
.08†
.00
Note. N=41 teams, Regression coefficients are standardized betas, cred = credibility.
**p<0.01,*p<0.05, †p<0.10
Table 13. Post-hoc Hierarchical Regression Analyses: TMS credibility and team outcomes
101
Team viability
Predictor
Task interdependence
Model 1
-.03
Model 2
-.03
Team satisfaction
Model 3
-.05
Model 1
Model 2
Model 3
.26*
.28*
.29*
Team reward
.31*
.31*
.34*
.10
.15
.15
Organizational support
.57**
.57**
.53**
.59**
.61**
.61**
.00
.03
TMS cred
TMS cred squared
R2
∆R2
-.11
.12
.51**
.51**
.52**
.00
.01
-.11
-.01
.50**
.51**
.51**
.01
.00
Note. N=41 teams, Regression coefficients are standardized betas, cred = credibility.
**p<0.01, *p<0.05, †p<0.10
Table 13. Continued
102
Team performance
Predictor
Team tenure
Team size
Country
Model 2
Model 3
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
.11
.08
.08
.29†
.28
.28
.21
.18
.18
-.27†
-.24†
-.24†
-.13
-.12
-.12
-.41**
-.37**
-.38**
.55**
.54**
.26
.31
.32
.23
.43*
.42*
.31†
.28
.10
.12
.34*
.31†
.37*
TMM squared
∆R2
Team creativity
Model 1
TMM
R2
Team learning
-.06
.32**
.38**
.39**
.06†
.01
.05
.27*
.28*
.28*
.01
.00
.05
.38**
.46**
.47**
.08*
.01
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 14. Results of Hierarchical Regression Analyses testing Hypothesis 1 with team tenure, team size and country of origin as
control variables: TMM Alternate scale
103
Team viability
Predictor
Team satisfaction
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Team tenure
-.04
-.10
-.10
.17
.11
.11
Team size
-.12
-.05
-.05
-.07
.00
.00
Country
-.12
.33†
.36†
-.18
.26
.27
.77**
.81**
.74**
.76**
TMM
TMM squared
R2
∆R2
.10
.03
.45**
.46**
.42**
.01
.04
.03
.43**
.43**
.40**
.00
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 14. Continued
104
Team performance
Predictor
Team learning
Team creativity
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
Task interdependence
.11
.09
.08
.00
-.03
-.05
.12
.10
.12
Team reward
.28
.35†
.31
-.01
.11
.08
.25
.33†
.38†
Organizational support
.19
.33†
.34†
.21
.46*
.46*
.18
.35†
.34†
-.44*
-.48*
TMM
-.26
TMM squared
R2
∆R2
-.30
-.09
.19*
.22*
.23*
.03
.01
-.30
-.08
.04
.15
.15
.11
.00
-.23
.16
.16†
.21†
.23†
.05
.02
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 15. Results of Hierarchical Regression Analyses testing Hypothesis 1 with task interdependence, team reward and
organizational support as controls: TMM Alternate scale
105
Team viability
Predictor
Task interdependence
Model 1
-.03
Team satisfaction
Model 2
Model 3
Model 1
Model 2
Model 3
.00
.02
.26*
.30*
.31*
Team reward
.31*
.22
.26†
.10
.00
.04
Organizational support
.57**
.37*
.36*
.59**
.37*
.36*
.36*
.41*
.40*
.44*
TMM
TMM squared
R2
∆R2
.12
.51**
.58**
.59**
.07*
.01
.10
.50**
.58**
.59**
.08*
.01
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 15. Continued
106
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11*
.09
.03
.03
.00
-.03
-.06
-.06
.12
.10
.04
.02
Team reward
.28
.35
.21
.21
-.01
.11
.05
.06
.25
.33†
.18
.19
Organizational support
.19
.33
.21
.21
.21
.46*
.41†
.41†
.18
.35†
.23
.22
-.26
-.23
-.20
-.44*
-.43*
-.41†
-.30
-.27
-.19
.33
.34
.13
.14
TMM
TMS
TMM X TMS
R2
∆R2
.10
.19*
.22*
.28*
.29*
.03
.06†
.01
.34†
.09
.04
.15
.16
.16
.11*
.01
.00
.36†
.31*
.16†
.21†
.27*
.36*
.05
.06†
.09*
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 16. Results of Hierarchical Regression Analyses examining impact of TMS on team outcomes: TMM Alternate Scale
107
Team viability
Predictor
Team satisfaction
Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
Task interdependence
-.03
.00
-.02
-.03
.26*
.30*
.30*
.29*
Team reward
.31*
.22
.16
.17
.10
.00
.00
.00
Organizational support
.57**
.37*
.32*
.32*
.59**
.37*
.37*
.37*
.36*
.37*
.40*
.40*
.40*
.43**
.13
.14
.01
.02
TMM
TMS
TMM X TMS
R2
∆R2
.11
.51**
.58**
.59**
.60**
.07*
.01
.01
.12
.50**
.58**
.58**
.59**
.08*
.00
.01
Note. N=41 teams, Regression coefficients are standardized betas.
**p<0.01, *p<0.05, †p<0.10
Table 16. Continued
108
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11
.09
.07
.06
.00
-.03
-.04
-.06
.12
.10
.09
.06
Team reward
.28
.35†
.22
.16
-.01
.11
.04
-.02
.25
.33†
.25
.12
Organizational support
.19
.33†
.30
.31
.21
.46*
.44*
.18
.35†
.33†
.34†
-.24
-.17
-.44*
-.43*
-.36
.27
.40†
.14
.28
TMM
-.26
TMS spec
TMM X TMS spec
R2
∆R2
.20
.19
.22
.28
.30
.03
.06
.02
.44*
-.30
-.29
.17
.22
.04
.15
.16
.19
.11*
.02
.03
-.14
.45*
.45*
.16†
.21†
.23†
.35*
.05†
.02
.12*
Note. N=41 teams, Regression coefficients are standardized betas, spec = specialization.
**p<0.01, *p<0.05, †p<0.10
Table 17. Results of Hierarchical Regression Analyses testing Hypothesis 2a and 3: TMM Alternate Scale
109
Team viability
Predictor
Model Model Model Model Model Model Model Model
1
Task interdependence
Team satisfaction
-.03
2
3
4
1
2
3
4
.00
.00
.00
.26*
.30*
.30*
.30*
Team reward
.31*
.22
.23
.21
.10
.00
.03
.00
Organizational support
.57**
.37*
.36*
.37*
.59**
.37*
.38*
.38*
.36*
.36*
.37*
.40*
.39*
.42*
TMM
TMS spec
-.03
TMM X TMS spec
R2
∆R2
.01
-.06
.06
.51**
.58**
.58**
.58**
.07*
.00
.00
-.01
.09
.50**
.58**
.58**
.59**
.08*
.00
.01
Note. N=41 teams, Regression coefficients are standardized betas, spec = specialization.
**p<0.01, *p<0.05, †p<0.10
Table 17. Continued
110
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11
.09
.09
.08
.00
-.03
-.01
-.04
.12
.10
.06
.05
Team reward
.28
.35†
.35
.35†
-.01
.11
.09
.08
.25
.33†
.35†
.35†
Organizational support
.19
.33†
.32
.31
.21
.46*
.51*
.46*
.18
.35†
.26
.25
-.26
-.28
-.30
-.30
-.41†
-.43†
.05
.07
.24
.27
TMM
TMS coord
TMM X TMS coord
R2
∆R2
-.44*
-.38
-.48†
.14
-.02
-.04
.19*
.22*
.22*
.23*
.03
.00
.00
-.17
.04
.15
.16
.18
.11*
.01
.02
-.04
.16†
.21*
.24*
.24*
.05†
.03
.00
Note. N=41 teams, Regression coefficients are standardized betas, coord = coordination.
**p<0.01, *p<0.05, †p<0.10
Table 18. Results of Hierarchical Regression Analyses testing Hypothesis 2b and 3: TMM Alternate Scale
111
Team viability
Predictor
Team satisfaction
Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
Task interdependence
-.03
.00
-.03
.02
.26*
.30*
.28*
.32*
Team reward
.31*
.22
.23†
.25†
.10
.00
.01
.02
Organizational support
.57**
.37*
.31*
.38*
.59**
.37*
.35*
.41*
.36*
.29†
.44*
.40*
.37*
.48*
TMM
TMS coord
.16
TMM X TMS coord
R2
∆R2
-.03
.07
.27*
.51**
.58**
.59**
.64**
.07*
.01
.05*
-.08
.20
.50**
.58**
.58**
.61**
.08*
.00
.03
Note. N=41 teams, Regression coefficients are standardized betas, coord = coordination.
**p<0.01, *p<0.05, †p<0.10
Table 18. Continued
112
Team performance
Predictor
Team learning
Team creativity
Model Model Model Model Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
1
2
3
4
Task interdependence
.11
.09
.05
.04
.00
-.03
-.06
-.07
.12
.10
.06
.02
Team reward
.28
.35†
.20
.23
-.01
.11
.01
.04
.25
.33†
.17
.23
Organizational support
.19
.33†
.24
.23
.21
.46*
.40†
.39†
.18
.35†
.25
.25
-.26
-.14
-.11
-.30
-.18
-.12
.27
.26
.28
.26
TMM
TMS cred
TMM X TMS cred
R2
∆R2
-.44*
-.37
-.34
.17
.16
.18
.19*
.22*
.27*
.29*
.03
.05*
.02
.17
.04
.15
.17
.19
.11*
.02
.02
.35*
.16†
.21†
.26*
.37*
.05†
.05†
.11*
Note. N=41 teams, Regression coefficients are standardized betas, cred = credibility.
**p<0.01, *p<0.05, †p<0.10
Table 19. Results of Hierarchical Regression Analyses testing Hypothesis 2c and 3: TMM Alternate Scale
113
Team viability
Predictor
Team satisfaction
Model Model Model Model Model Model Model Model
1
2
3
4
1
2
3
4
Task interdependence
-.03
.00
-.02
-.03
.26*
.30*
.29*
.29*
Team reward
.31*
.22
.13
.14
.10
.00
-.02
-.01
Organizational support
.57**
.37*
.31*
.31*
.59**
.37*
.36*
.36*
.36*
.43*
.43*
.40*
.41*
.42*
.15
.15
.04
.03
TMM
TMS cred
TMM X TMS cred
R2
∆R2
.03
.51**
.58**
.59**
.59**
.07*
.01
.00
.07
.50**
.58**
.58**
.59**
.08*
.00
.01
Note. N=41 teams, Regression coefficients are standardized betas, cred = credibility.
**p<0.01, *p<0.05, †p<0.10
Table 19. Continued
114
115
TMM AD by TMS
1.00
0.80
0.60
Team creativity
0.40
0.20
High TMM AD
Average TMM AD
0.00
-0.20
TMS
TMS
TMS
Low
Average
High
-0.40
-0.60
-0.80
-1.00
Figure 3. Interaction between TMM and TMS on team creativity
Low TMM AD
116
Team creativity by TMS Specialization
1.00
0.80
0.60
Team creativity
0.40
0.20
0.00
-0.20
TMS Spec
TMS Spec
TMS Spec
Low
Average
High
-0.40
-0.60
-0.80
-1.00
Figure 4. Curvilinear impact of TMS specialization on team creativity
117
Team viability by TMS Coordination
1.00
0.80
0.60
Team viability
0.40
0.20
0.00
-0.20
TMS Coord
TMS Coord
TMS Coord
Low
Average
High
-0.40
-0.60
-0.80
-1.00
Figure 5. Curvilinear impact of TMS coordination on team viability
118
Team satisfaction by TMS Coordination
1.00
0.80
0.60
Team satisfaction
0.40
0.20
0.00
-0.20
TMS Coord
TMS Coord
TMS Coord
Low
Average
High
-0.40
-0.60
-0.80
-1.00
Figure 6. Curvilinear impact of TMS coordination on team satisfaction
119
Team performance by TMS Credibility
1.00
0.80
0.60
Team performance
0.40
0.20
0.00
-0.20
TMS Cred
TMS Cred
TMS Cred
Low
Average
High
-0.40
-0.60
-0.80
-1.00
Figure 7. Curvilinear impact of TMS credibility on team performance
120
TMM AD by TMS Specialization
1.00
0.80
0.60
Team creativity
0.40
0.20
High TMM AD
Average TMM AD
0.00
-0.20
TMS Spec
TMS Spec
TMS Spec
Low
Average
High
Low TMM AD
-0.40
-0.60
-0.80
-1.00
Figure 8. Interaction between TMM and TMS specialization on team creativity
121
TMM AD by TMS Coordination
1.00
0.80
0.60
Team performance
0.40
0.20
High TMM AD
Average TMM AD
0.00
-0.20
TMS Coord
TMS Coord
TMS Coord
Low
Average
High
Low TMM AD
-0.40
-0.60
-0.80
-1.00
Figure 9. Interaction between TMM and TMS coordination on team performance
122
TMM AD by TMS Credibility
1.00
0.80
0.60
Team Performance
0.40
0.20
High TMM AD
Average TMM AD
0.00
-0.20
TMS Cred
TMS Cred
TMS Cred
Low
Average
High
Low TMM AD
-0.40
-0.60
-0.80
-1.00
Figure 10. Interaction between TMM and TMS credibility on team performance
123
TMM AD by TMS Credibility
1.00
0.80
0.60
Team learning
0.40
0.20
High TMM AD
Average TMM AD
0.00
-0.20
TMS Cred
TMS Cred
TMS Cred
Low
Average
High
Low TMM AD
-0.40
-0.60
-0.80
-1.00
Figure 11. Interaction between TMM and TMS credibility on team learning
124
TMM AD by TMS Credibility
1.00
0.80
0.60
Team creativity
0.40
0.20
High TMM AD
Average TMM AD
0.00
-0.20
TMS Cred
TMS Cred
TMS Cred
Low
Average
High
Low TMM AD
-0.40
-0.60
-0.80
-1.00
Figure 12. Interaction between TMM and TMS credibility on team creativity
125
TMM Alt by TMS
1.00
0.80
0.60
Team creativity
0.40
0.20
High TMM alt
Average TMM alt
0.00
-0.20
TMS
TMS
TMS
Low
Average
High
-0.40
-0.60
-0.80
-1.00
Figure 13. Interaction between TMM and TMS on team creativity
Low TMM alt
126
TMM Alt by TMS Specialization
1.00
0.80
0.60
Team creativity
0.40
0.20
High TMM alt
Average TMM alt
0.00
-0.20
TMS Spec
TMS Spec
TMS Spec
Low
Average
High
Low TMM alt
-0.40
-0.60
-0.80
-1.00
Figure 14. Interaction between TMM and TMS specialization on team creativity
127
TMM Alt by TMS Coordination
0.90
Team viability
0.40
High TMM alt
Average TMM alt
-0.10
TMS Coord
TMS Coord
TMS Coord
Low
Average
High
Low TMM alt
-0.60
-1.10
Figure 15. Interaction between TMM and TMS coordination on team viability
128
TMM Alt by TMS Credibility
1.00
0.80
0.60
Team creativity
0.40
0.20
High TMM alt
Average TMM alt
0.00
-0.20
TMS Cred
TMS Cred
TMS Cred
Low
Average
High
Low TMM alt
-0.40
-0.60
-0.80
-1.00
Figure 16. Interaction between TMM and TMS credibility on team creativity
129
Key findings
TMM (AD)
TMM
(6-item scale)
Hypothesized relationships
TMM has curvilinear relationship with team
performance (H1).
TMS specialization is positively related to team
performance (H2a).
TMS coordination is positively related to team
performance (H2b).
TMS credibility is positively related to team
performance (H2c).
TMM moderates relationship between TMS and
team performance (H3).
Not supported
Not supported
Not supported
Not supported
Not supported
Not supported
Partially
supported
Not supported
Partially
supported
Not supported
Post-hoc Analysis
TMS specialization exhibits curvilinear
relationship with team creativity.
TMS coordination exhibits curvilinear
relationship with team viability.
TMS coordination exhibits curvilinear
relationship with team satisfaction.
TMS credibility exhibits curvilinear relationship
with team performance.
TMM moderates TMS and team creativity
Figure 3a
Figure 4a
Figure 5a
Figure 6a
Supported
(Figure 4)
Supported
(Figure 8)
Supported
(Figure 9)
Supported
(Figure 10)
Supported
(Figure 11)
Not supported
Supported
(Figure 13)
Supported
(Figure 14)
Not supported
TMM moderates TMS specialization and team
creativity
TMM moderates TMS coordination and team
performance
TMM moderates TMS credibility and team
Not supported
performance
TMM moderates TMS credibility and team
Not supported
learning
TMM moderates TMS coordination and team
Supported
viability
(Figure 15)
TMM moderates TMS credibility and team
Supported
Supported
creativity
(Figure 12)
(Figure 16)
a
Note. Relationships are independent of method of TMM operationalization.
Table 20. Summary of key findings.
130
CHAPTER 5: DISCUSSION
The purpose of the present study was to examine the contribution of two
constructs – TMM and TMS, in predicting important team effectiveness dimensions
(team performance, team learning, team creativity, team viability and team satisfaction)
in field settings. Building on previous research, I was interested in examining the
relationship between transactive memory systems (TMS) and team mental models
(TMM). In addition, I explored the impact of TMS components (specialization,
coordination and credibility) on team effectiveness dimensions. Further, I explored
whether two different methods of operationalizing TMM (i.e. average deviation method
vs. likert scale) would yield different results.
Key findings can be summarized under three broad headings: 1) TMM, 2) TMS
and 3) Interaction between TMM and TMS. First, the relationship between TMM and
team performance in a field setting appears to be more complex than a simple linear
relationship observed in previous studies undertaken in laboratory settings. In addition,
TMM failed to predict team learning or team creativity. Even the hypothesized
curvilinear effects between TMM and team performance received no support. However,
TMM had an impact on team viability and team satisfaction. In addition, the two very
different and unrelated methods to compute TMM agreement (average deviation and the
6-item scale) led to similar patterns of relationships, suggesting both methods were
capturing something similar even though they are virtually uncorrelated. Given that the
average deviation indices performed as well as the alternative 6-item scale, this study
provided initial support for using the average deviation technique as a conceptually and
methodologically valid tool to measure TMM in a field setting. Hopefully, this may spur
more studies on TMM using “real-life” organizational teams.
131
Second, TMS deserves to be treated as a multidimensional construct since the
narrow specific dimensions are more helpful and predictive of team outcomes than the
omnibus TMS construct. In particular, TMS specialization and TMS credibility were
stronger indicators of the higher order TMS construct than TMS coordination. This is
consistent with the findings that TMS specialization and TMS credibility are correlated,
but neither correlates with TMS coordination, indicating that TMS coordination behaves
differently than other components. Moreover, the hypothesized linear effect between the
TMS components (specialization and coordination) and team performance was not
supported. In particular, TMS credibility seemed to drive team performance and team
creativity while TMS coordination was predictive of team viability and satisfaction. The
results of post-hoc analyses suggested the presence of a non-linear relationship between
the TMS components (specialization, coordination and credibility) and team effectiveness
dimensions (e.g. team performance and creativity).
Third, no moderation effect of TMM was observed for the relationship between
TMS and team performance. However, TMM was found to moderate the relationship
between TMS and team creativity. Again, the interaction patterns of TMM on team
effectiveness outcomes were stronger when specific TMS components were examined.
Interestingly, the pattern of interaction between various TMS components and team
effectiveness outcomes was strikingly similar. In general, the interaction pattern
suggested a compensatory relationship between TMM and TMS components. Whenever
a team experienced a low TMM, a high level of TMS component (say credibility) helped
overcome the detrimental effects of low TMM to achieve high team performance. In
contrast, a high level of TMM helped offset the downside of a low TMS component (e.g.
132
credibility) on team performance. Contrary to expectations, a high level of TMM and
TMS component did not result in the most optimal team performance (or creativity).
Further, the results were mostly consistent across both methods of
operationalizing TMM (average deviation vs. likert scale), although there were some
minor differences. Specifically, TMM moderated the relationship between TMS
credibility and team performance under the average deviation method but not under the
Likert method. A similar result was observed with team learning under the average
deviation index but not under the alternative method. Further details of the differences
were discussed in the results section. Next, I discuss the theoretical and managerial
implications of these results.
Theoretical Implications
Team learning
In a recently published article in The Academy of Management Annals,
Edmondson, Dillon and Roloff (2008) commented that “team learning…is an useful
rubric, an umbrella term encompassing a variety of loosely related theories and studies”
(p.302). Consistent with information processing theory, the authors further suggest that
terms such as Shared Mental Models and Transactive Memory Systems are conceptually
similar since they are all team level cognitive systems that encode, store, retrieve and
communicate knowledge, and are used to predict task performance. Yet surprisingly,
there are few studies that investigate all of these similar sounding constructs. The present
study is an early attempt to help bridge the research carried out in experimental settings
(transactive memory systems, shared mental models) with those in field research
(subsumed under group learning or team learning). This work addresses a key issue
133
raised by many scholars who state the necessity to integrate existing findings (Argote,
Gruenfeld, & Naquin, 2001; Mohammed & Dumville, 2001).
Extending on Ellis et al. (2006), I proposed that SMM and TMS are two
mechanisms of team learning, and I examined the relationship between these two
constructs. I attempted to conceptually clarify the construct of SMM as having an
integrative function and the construct of TMS as having a differentiating function. In
addition, I integrated the literature on information sampling (Stasser, Taylor, & Hanna,
1989; Stasser, Vaughan, & Stewart, 2000), information processing (Hinsz, Tindale, &
Vollrath, 1997) and organizational team research (Argote, Gruenfeld, & Naquin, 2001) to
extend the relationship between SMM and TMS. Table 21 reflects the conceptual
overlap.
Perspective
TMM
TMS
Present study
Integrative
Differentiation
Group learning
Interpersonal focus
Task focus
Information sampling
Shared information
Unshared information
Information processing
Sharing
Storage and Retrieval
Table 21. Conceptual overlap across different literatures on team learning.
134
As seen in Table 21, the current study clarifies how TMM and TMS are
conceptually similar to constructs discussed in various literatures related to team learning.
First, the group learning literature stresses the importance to acquire, share, refine, or
combine task-relevant knowledge through interaction with one another (Argote,
Gruenfeld, & Naquin, 2001). It is clear that such a definition has two important
components – a) interpersonal interaction and b) task-relevant knowledge. A TMM is a
measure of interpersonal awareness and adequately captures the sharing of information
across team members while TMS with its differentiation focus captures how task-relevant
knowledge is distributed.
Second, small group researchers working mostly in experimental settings have
developed an amazing literature on how groups work. This literature has been commonly
referred to as the information sampling framework. A common result of their study is that
groups work best when they combine all the relevant information, shared and unique,
dispersed and available, to their members (Stasser, Vaughan, & Stewart, 2000; D. D.
Stewart & Stasser, 1995). It has been suggested that groups are able to solve the most
complex problems when they are able to incorporate unique information from out-group
members or strangers (Gruenfeld, Mannix, Williams, & Neale, 1996; Thomas-Hunt,
Ogden, & Neale, 2003). This is consistent with the present theoretical conceptualization
that we need a match between TMM, focusing on common knowledge and interaction,
and TMS, focusing on unique knowledge dispersed within the group.
Third, the information sharing framework suggests that groups act as information
processors when they efficiently share, store and retrieve information (Hinsz, Tindale, &
Vollrath, 1997). Recently, Wilson et al. (2007) also characterized group learning as a
135
three step process of sharing, storage and retrieval. I propose that TMM focuses on the
sharing process while TMS captures the storage and retrieval part. Hence, the present
conceptualization of the TMM and TMS constructs is consistent with those enumerated
by the information-processing framework.
A major limitation of the previous work has been the lack of clarity about whether
scholars are studying team learning as an outcome or as a process (Wilson, Goodman, &
Cronin, 2007). I propose that team learning processes are explained by the constructs of
SMM and TMS, while learning outcomes are better explained by team learning behaviors
as observed by team managers. This is consistent with the argument that research focused
on task mastery jobs (e.g. SMM and TMS) in experimental settings has been focusing on
team learning processes while scholars investigating group processes have focused on
team learning outcomes (Edmondson, Dillon, & Roloff, 2008).
A major contribution of this study is to integrate the research on team learning
from multiple perspectives by making an explicit assumption that TMM and TMS are
two emergent learning processes and are different from the construct of team learning as
an outcome. I tested this assumption by explicitly measuring TMM, TMS and team
learning. Another related contribution of this study is to extend the idea that TMM
captures the integrative function while TMS focuses on the differentiation aspects within
a team.
Finally, I delineated the team learning construct from team performance as
Wilson et al. (2007) argue in the following statements that team learning is not equivalent
to team performance as implicitly assumed (p.1043).
We argue that learning may have occurred, even when there was
no change in a group’s overall performance. For example, the
136
group may have learned something but may not have had an
opportunity to apply the learning in a way that would change its
performance. Conversely, performance can change without any
learning actually taking place.….. Finally, learning does not
always result in positive outcomes. Research on group learning
needs to account for the possibility of dysfunctional learning, as in
the case of superstitious learning, where a group learns a false
connection between its actions and some outcomes.
In the present study, team learning and team performance were highly correlated
suggesting that learning was integral to better team performance. Yet, the results clearly
suggest that team learning is quite distinct from team performance. It is important to
understand that the predictors of team performance may not predict team learning. In
addition, team learning and team creativity were quite distinct. Moreover, the present
study highlights that the importance of explicitly considering the dependent variables of
interest. The results suggest that team performance was contingent on TMS credibility
but not on either TMS specialization or TMS coordination. In other words, the belief or
trust in the accuracy of the teammates matters most for high team performance. Further,
the actual distribution of knowledge amongst teammates, and the ability to work together,
does not predict team performance. Similarly, team creativity was dependent on TMS
credibility and TMS specialization, but not on TMS coordination. Thus, team creativity
does not depend on team members’ ability to work together.
Further, team coordination predicted team viability and team satisfaction. My
study provides empirical support to the notion that team learning and team performance
are not interchangeable, and are two distinct constructs with unique predictors. Hence,
the idea that engagement in team learning will always lead to better team performance or
vice-versa is not a tenable argument.
137
Context
As anticipated, contextual variables like team size, team tenure and country of
origin were strong predictors of team effectiveness measures. It was interesting to
observe that team tenure was related to increased team learning suggesting that team
members may need to work together for a substantial time before they start engaging in
team learning. This may be because team members need to know each others’
preferences and skills/ beliefs before coming together to find common ground on how to
change their established behaviors.
This is consistent with the findings of Gladstein (1984) that team performance in
organizational settings is strongly dependent on organizational features. It is possible that
there is an optimum team size beyond which team performance and team creativity suffer
due to problems in managing coordination within the team.
In my study, teams from the US displayed higher team performance, engaged in
more learning behaviors and team creativity compared to teams from India. It should be
mentioned that almost all teams from India were from the IT industry, while the teams
from the US were more diverse representing multiple industries. It is likely that teams
from the IT industry behave differently compared to those in other industries. Hence, it is
hard to pinpoint whether the observed differences were due to national or industry
characteristics. The results underscore the importance of considering contextual factors
like geography and organizational characteristics as boundary conditions for the
applicability of theoretical models in team effectiveness (Johns, 2006).
Contextual variables like organizational support and team reward helped increase
team viability. It appears that goals associated with team rewards help accentuate team
138
members’ social identity with self and others into in-group and out-group. Selfcategorization theory predicts that a superordinate goal applicable to everyone in the team
(e.g. team rewards) will make the team work together for achieving the goal, resulting in
relatively stable group structure and increased team viability (Hogg & Terry, 2000).
Organizational support also helped increase team members’ satisfaction with each other.
These results reiterate the importance of considering contextual variables in order
to predict team level outcomes. Laboratory experiments generally fail to capture the
intricacies of teams in professional settings. In my sample comprising of teams from
diverse backgrounds, contextual variables (e.g. team size, team tenure, country of origin,
organizational support) were strong predictors of team effectiveness dimensions. Ignoring
these important variables that are vital in any field study of “real-life” organizational
teams will provide an incomplete picture of the actual team processes.
TMM
In particular, this study using the Average Deviation Index of TMM failed to
support the notion that Mental Models had a linear effect on team performance, a result
inconsistent with those obtained by Mathieu and colleagues (Mathieu, Goodwin, Heffner,
Salas, & Cannon-Bowers, 2000; Mathieu, Heffner, Goodwin, Cannon-Bowers, & Salas,
2005). However, the present research examined organizational work teams whereas
Mathieu and colleagues have looked at mainly student samples. Also, the contextual
variables provided a stronger effect on team performance and could be a reason why
Team Mental Model (TMM) failed to predict team performance. However, the results are
consistent with a rare field study conducted on TMM in a field setting which failed to
139
observe linear relationships between TMM and team performance (Smith-Jentsch,
Mathieu, & Kraiger, 2005).
Operationalizing TMM
The operationalization of TMM as a 6-item scale was supportive of the linear
relationships between TMM and team performance. Since the proposed model was not
validated by one of the two operationalizations of TMM, there is a need for further tests
to investigate the inconsistent results obtained in this study in more field settings. It is
clear that choosing the method to measure TMM has an impact on the results and
implications for researchers working in this area.
The results from exploring average deviation indices (Burke & Dunlap, 2002) and
using a six-item alternative likert scale (Appendix B) as indices of TMM were similar in
one respect yet different in many ways. The results yielded similar results in relation to
the three hypothesized relationships; yet, the post-hoc results yielded two different
patterns and a slightly different explanation suggesting that computing agreement by rank
ordering of the team processes by team members leads to an effective measure of TMM.
The average deviation captures the agreement level in a team and hence seems to be an
effective measurement method.
TMS and team effectiveness
TMS was more helpful than TMM in predicting team performance. Interestingly,
the main component of TMS driving this relationship appears to be TMS credibility. In
addition, TMS credibility also predicted team creativity. The results suggest that teams
are more creative and better performers when members trust each other’s expertise. This
140
raises the suggestion that it is vital to develop recognition and credibility around the
expertise available within teams.
The exploration of the curvilinear relationships between TMS components and the
various team effectiveness dimensions led to some interesting results. It was important to
highlight that different TMS components behaved differently with the dependent
variables of interest. Specifically, TMS credibility exhibited an increasing hyperbolic
relationship with team performance, TMS specialization exhibited a U-shaped
relationship with creativity and TMS coordination exhibited the same with team viability
and satisfaction. In general, team performance was best under the highest levels of TMS
credibility, suggesting that it is important for team members to identify and trust the
expertise available within teams. A high TMS credibility implies implicit trust about the
accuracy of other members’ knowledge. In contrast, a low TMS credibility suggests that
members do not trust each other’s expertise, and this may lead each member to waste
time and effort by cross-checking the information received. In terms of informationprocessing theory, it is possible that when teammates have strong belief in each other,
they will spend less time searching for information. This may lead to efficient retrieval of
information, which should result in better team performance. When team members trust
each other they are more likely to try out new ideas suggested by their members leading
to higher team creativity.
Similarly, results suggest that teams high on TMS specialization engage in more
creative behavior. This is likely because teams high on specialization will be comprised
of individuals with a high level of knowledge differentiation. A team with different
viewpoints is more likely to engage in new ways to do things and be creative. Another set
141
of results that TMS coordination predicts team viability and team satisfaction has a
simple explanation. TMS coordination refers to team members’ ability to work together
efficiently. Thus, teams are more likely to continue working together in the future and
engage in behaviors essential to maintaining itself for future tasks or be viable. In
addition, team members are more likely to be satisfied with each other when they can
work together efficiently.
TMS and TMM interaction
When I examined the interactions between TMM and TMS, the two different
operationalizations of TMM led to different results. The one clear pattern from
examining the results was that measuring TMM in terms of average deviation led to
superior results in terms of the observed interactions amongst TMM and TMS. This
suggests that it is important to measure agreement amongst team members about their
teammates and team processes. I’ll first discuss the results obtained when TMM was
operationalized as an average deviation index.
Contrary to expectations, TMM and TMS did not have an additive effect on team
performance and team learning but more of a compensatory effect. High TMM made up
for low TMS. As expected, TMM moderated the relationship between TMS and team
creativity but in an unexpected way. As expected, teams had the lowest creativity when
both TMS and TMM were lowest. Surprisingly, results revealed that team creativity was
highest under conditions of low TMM and high TMS. The results suggest lack of TMM
frees team members from a rigid mindset of doing things, unleashing more options and
fostering the climate for innovation and creativity. In addition, my results suggest a
consistent pattern of moderation between TMS credibility and TMM on team
142
performance, team learning and team creativity. A similar pattern of results was also
observed between TMS specialization and team creativity.
One reason why we observe the compensatory effect of TMM and TMS in
predicting team performance and team learning could be that a high degree of TMM
coupled with high trust levels regarding expertise in a team (TMS credibility) builds a
sense of complacency within the team. It is also possible that in such situations, project
teams are expected to bring their different expertise and viewpoints towards completion
of organizational tasks. However, teammates look toward their “experts” and fail to
engage in an open ended constructive dialogue critical for effective team performance
(Tjosvold, 1985). This may result in the building of unusually high trust within teams,
which may lead to stifled creativity as a potential downside (Wicks, Berman, & Jones,
1999). Thus, high TMM and TMS credibility lead to a situation where teams engage in
less idea-generation, a negative consequence of both high TMM and high TMS
credibility. This may be partially responsible for the absence of team learning behaviors
because the team has a mental model that suggests a dysfunctional level of agreement.
The results were slightly different when measured with a 6-item scale to measure
TMM. TMM moderated the relationship between TMS specialization and team creativity
as observed before, but the pattern of interaction was slightly different (Figure 14) than
that observed when TMM was measured as an average deviation index (Figure 8). Teams
displayed the highest creativity levels when they demonstrated both high TMS
specialization and high levels of TMM. Conversely, teams exhibited the lowest creativity
under conditions of high TMM and low TMS specialization. It appears that high levels of
TMM are unhelpful and lead team members to behave in an established way of doing
143
things. This condition of high TMM seems especially problematic when there is low
TMS specialization, or a lack of specialists. A team with a high level of shared
knowledge, low TMS specialization, is hardly going to be creative when they all share a
high TMM. On the contrary, TMM is most helpful when team members focus on their
own task and engage in sharing the unique knowledge dispersed among team members.
The pattern of interaction observed between TMS credibility and TMM on team
creativity is similar. Under conditions of low TMS credibility when team members have
very little trust in each other, a high level of TMM was particularly disruptive. Naturally,
team creativity was lowest when teammates trusted each other the least and were under a
situation of high TMM. Teams did best when they had high TMM and a strong belief in
each other, as evident by high TMS credibility.
As expected, team viability was highest under conditions of high TMM and high
TMM coordination. Stated another way, when teammates have an ability to work
together efficiently, as captured by the notion of high TMS coordination and high
knowledge about teammate’s characteristics and their interpersonal interaction patterns,
captured by high TMM, they express interest in continuing to work together. In contrast
when TMM is low, teammates are wary of their teammate’s reaction to a particularly
novel situation that may arise in due course. Teams exhibit low willingness to continue
together when they have teammates who have the ability to work together but don’t do
so. This might create an impression that it is better to dissolve the existing team than
continue working together.
It was interesting to observe a moderate correlation between TMS coordination
and the alternative 6-item TMM scale. It appears that there is a similar reaction towards
144
TMS coordination as towards TMM. This is expected, as per the definition of these two
constructs. If there is substantial agreement amongst team members, it is quite likely that
they will have a high TMM as well as believe in their ability to work together efficiently
leading to high TMS coordination. In some ways, TMS coordination has the closest
linkages to TMM, suggesting that these two constructs may be capturing the same
underlying phenomenon. However, the lack of a similar relationship with the average
deviation method does weaken the argument.
Managerial Implications
The results provide helpful tips for practitioners interested in team learning. While
many practitioner oriented books have suggested that team learning is very important, it
has been rather difficult to implement the suggestions in practice. I hypothesized that
TMM and TMS are important antecedents to team performance. The results illustrate that
contextual factors like team size, team tenure and country are strong predictors of team
performance and team learning. Managers should consider the impact of increasing team
size in organizations. While more people in a team helps bring multiple and often varied
expertise to a team, managing divergent viewpoints and conflict becomes an important
drawback resulting in lower than expected team performance (Haleblian & Finkelstein,
1993). This is especially important as organizational work teams have been increasing in
team size over the years. The fact that increased team tenure helps increase team learning
is critical for organizations that are promoting intellectual capital. The fact that team
learning is an added benefit of retention will be of interest to human resource executives.
The cross-cultural study also highlights the importance of considering the influence of
national work cultures. A majority of the teams were from organizations with a
145
substantial multinational presence, and yet the results suggested that teams from the US
had better team performance than those from India.
Results suggest that TMS credibility plays a key role in improved team
performance and team creativity. The results are consistent with prior findings that team
processes are more important for team effectiveness than individual members’
competence when teams work together for a long time (Watson, Michaelsen, & Sharp,
1991). This study supports the conventional wisdom that it is important to build trust in
team members’ knowledge rather than staffing a team with individual star performers
(Gladwell, 2002).
Organizations can help teams achieve high TMS credibility by identifying experts
and encouraging team members to believe in each other by assigning tasks that build trust
and reliance on each other. Additionally, results demonstrate that it is important to
determine team objectives before engaging in any team-building activities. Managers
must decide whether they want to focus primarily on team performance, team creativity
or team viability. The choice of the specific objective must be matched with
corresponding TMS components. In my study, high performance teams were high on
TMS credibility, while the most satisfied teams were high on TMS coordination. This
implies that different mechanisms are at play in achieving desired team effectiveness
dimension.
Senior executives and organizations should also realize that spending time and
effort on improving TMM is not always a good thing (Senge, 1990) because teams in my
sample did perform well despite low levels of TMM. Managers should consider the
alternative approach of improving TMS credibility to achieve high team performance.
146
More importantly, managers must take into account the contingency approach between
balancing TMS and TMM. High team performance can be obtained by being high on one
construct and low on the other. Practitioners should take into account the match between
TMM and TMS before deciding on how teams should achieve the desired objectives. If
the team has high TMM, managers may not benefit from trying to improve TMS
mechanisms. On the other hand, if the team exhibits low TMM, efforts to improve TMS
mechanisms can be helpful. All of the preceding suggestions can be facilitated by
managers surveying team members in order to find out initial levels of TMM and TMS.
Limitations
The relatively small sample size in this study limited the statistical power to
detect significant relationships. The sample for the present study comprised of 41 teams,
and is similar to those in many other team level studies like a 45 team study reported by
Stewart and Barrick (2000) as well, average project teams studies reported by Cohen and
Bailey (1997). However, some strong results were observed between team performance,
team size and country. The presence of the interactive relationship between TMM and
TMS credibility suggests the presence of strong effects irrespective of the small sample
size. Thus, it is important to further explore these constructs to better understand team
performance and learning.
Second, the present sample may not be generalizable to all types of organizations.
However, the study sample consisted of 41 teams from 11 organizations across 7
industries from two countries. The multi-industry, multi-organization sample resulted in
increased variance in the sample. Hence, the results are expected to be conservative
147
estimates and should generalize better than prior studies focused on a particular industry
or an organization.
Third, the cross-sectional nature of the data limits the findings. Teams may have
been at different stages in their life-cycle, some newly formed, others mature and yet
others working on their last project together when the data were collected. The average
team tenure was 2.9 years, suggesting most teams had been working together for a
substantial length of time and the lack of supportive findings are not due to the survey
timing. It would have been beneficial if all the teams were at the same level of formation,
and the data collection was done longitudinally to better understand the team learning
processes. Hence, any inference of causality should not be deduced because of the
correlational nature of data.
Fourth, although many measures were collected from team managers, some
results may have been influenced by common method bias. In particular, common
method bias could have influenced the relationship between team viability and
satisfaction and TMM and/or TMS. The team processes (TMM and TMS), team viability,
and team satisfaction measures were collected from team members rather than external
observers.
Fifth, although I had a reasonably high response rate, 58% response within each
team, for a field study done remotely by online surveys, team level constructs such as
TMM and TMS are more representative when complete information from every team
member is available. However, it took multiple reminders to achieve the present response
rate and researchers should be mindful of the low response rate in field settings.
148
Sixth, I operationalized TMM as an average deviation index, a type of difference
score, following scholars studying the mental model literature (cf. Smith-Jentsch,
Mathieu, & Kraiger, 2005). The use of difference scores has been criticized by
researchers as this may not capture the agreement constructs like TMM effectively.
Edwards (Edwards, 1995; Edwards & Parry, 1993) has recommended that researchers use
other techniques like polynomial regression techniques to operationalize the agreement
constructs like TMM.
Seventh, many of the findings were of marginal significance (p≤ .10). This
increases the probability of finding results by chance, a direct result of Type I error.
Hence, these results should be interpreted with caution. Replication studies are needed to
support these results.
Finally, I chose to aggregate constructs to team level even when some agreement
indices were lower than the conventionally accepted cut-off norms. This may have
affected the results. It is possible that my sample represents a lower level of agreement
amongst team members than has been observed in the published research. This may
reflect team members having significant within-group differences. Future researchers
should carefully model agreement in future research differences, as well the similarities
amongst team members, by using a multilevel procedure (Kozlowski & Klein, 2000).
Fortunately, the chi-square tests under confirmatory factor analyses were significant
despite the small sample size, which provides a conservative estimate that items do
represent the underlying constructs.
149
Future Research
Future researchers could pursue many different avenues. First, it is important to
investigate the conditions under which teams engage in learning behaviors. In this study,
I did find that contextual variables play an important role in predicting team-level
outcomes, but was unable to pinpoint the impact of industry and the country of origin on
team effectiveness. Due to the paucity of field studies involving TMM and TMS, there is
a need to engage in more field studies to replicate the findings of this study. Scholars
should investigate multiple industries, and countries to examine their impact on teamlevel dimensions. It is also important to examine more proximal organizational design
features that may affect team performance and team learning.
Second, TMS and TMM are conceptualized as emergent processes that emerge
over a period of time. The current study was a cross-sectional study and included teams at
all levels of maturity. It will be helpful in future studies if teams are tracked from their
time of formation to the time of disbanding, and regular measurements at equal intervals
are done to track the development of TMS and TMM. Hence, it is important to do a
longitudinal study of the teams to get a better sense of how TMS and TMM develop, as
well as resulting impact on the team level outcomes like team performance, team learning
and creativity.
Third, there is an urgent need to clarify both conceptually and methodologically
the nature of the TMM construct (Mohammed, Klimoski, & Rentsch, 2000). Ironically,
unless there is an agreement amongst researchers working on the mental model construct,
the research is likely to remain fragmented and of little use to researchers and
practitioners alike. I conceptualized two different ways to capture TMM, and the similar
150
pattern across both methods suggests that the TMM construct can be meaningfully
captured using an average deviation index in field settings. This is an initial effort to
measure TMM in “real-life” teams and needs to be replicated in other settings.
Fourth, the finding that the TMS construct is better represented as a three-factor
structure under an umbrella construct deserves more investigation. The evidence that the
narrow dimensions (e.g. specialization, coordination and credibility) predict team
outcomes (e.g. team performance and team creativity) better than a broader construct is
helpful. In the future, researchers should include these three factors as stand-alone
predictors of variables such as team performance.
Fifth, the results suggest an extremely strong relationship between TMS, TMM
and team creativity. Team creativity turned out to be more strongly related to the TMS
components than team learning. This unexpected relationship can be utilized by
researchers focused on creativity, as TMS specialization and TMS credibility turned out
to be important predictors of team creativity.
Sixth, in order to integrate findings across similar sounding labels, it is important
that researchers take a broader perspective when they continue working in this field and
adopt a cross-fertilization approach to assimilate the existing findings and clarify the
meaning underlying the term team learning. The confusion regarding the construct is
quite evident by recent summaries of literature around this topic, and it is no wonder that
the existing research is so fragmented that researchers hope for more integration
(Edmondson, Dillon, & Roloff, 2008; Wilson, Goodman, & Cronin, 2007).
Seventh, it could help the team learning literature if researchers engage in a
systematic, objective review of existing studies and conduct a meta-analysis. A meta-
151
analysis might highlight the major findings and settle confusion in the existing literature.
It is important that we do not engage in the minor replication of existing studies, but
rather identify specific areas which can help advance the team learning and related
literature.
Finally, researchers working in this area should not only investigate the interrelationship between TMS and TMM but they should also incorporate related cognitive
constructs that are most likely related to the concept of team learning. Such terms might
include information sharing, cognitive consensus and group learning amongst other
similar sounding labels.
Conclusion
Despite multiple calls for integrating the research on various team learning related
constructs, there is little empirical research especially in field settings. In this study, I
tested the relationship between team mental models (TMM) and transactive memory
systems (TMS) and their impact on team outcomes in the field. Although the results
generally advance our understanding of these team learning mechanisms, the results
generally failed to support the existence of a positive relationship between TMM and
team performance. In addition, TMM failed to moderate the relationship between TMS
and team performance. On further evaluation of TMS components, the results were more
promising. TMS credibility was positively related to team performance and team
creativity. I found that TMM moderated the relationship between TMS credibility and
team outcomes. Additionally, high levels of TMM were found to be less than optimal for
effective team performance, especially when teams demonstrated high TMS. I hope that
my study will provide a stimulus for future research investigating team learning.
152
REFERENCES
Aiken, L. S., & West, S. G. (1991). Multiple Regression: Testing and
Interpreting Interactions. Newbury Park, CA: Sage.
Aldag, R. J., & Fuller, S. R. (1993). Beyond Fiasco: A Reappraisal of the
Groupthink Phenomenon and a New Model of Group Decision Processes.
Psychological Bulletin, 113(3), 533-552.
Amabile, T. M., Conti, R., Coon, H., Lazenby, J., & Herron, M. (1996).
Assessing the Work Environment for Creativity. The Academy of Management
Journal, 39(5), 1154-1184.
Ancona, D. G., & Caldwell, D. F. (1992). Bridging the Boundary:
External Activity and Performance in Organizational Teams. Administrative
Science Quarterly, 37(4), 634-665.
Argote, L., Gruenfeld, D., & Naquin, C. (2001). Group Learning in
Organizations. In M. E. Turner (Ed.), Groups at Work: Theory and Research (pp.
369-411). Mahwah, NJ: Lawrence Erlbaum Assoc Inc.
Argyris, C. A., & Schon, D. A. (1978). Organizational Learning. Reading,
MA: Addison-Wesley.
Austin, J. R. (2003). Transactive Memory in Organizational Groups: The
Effects of Content, Consensus, Specialization, and Accuracy on Group
Performance. Journal of Applied Psychology, 88(5), 866-878.
Bandura, A. (1977). Social Learning Theory. New York: General
Learning Press.
Bandura, A. (1986). Social Foundations of Thought and Action: A Social
Cognitive Theory. Englewood Cliffs, NJ: Prentice-Hall.
Blackler, F. (1995). Knowledge, Knowledge Work and Organizations: An
Overview and Interpretation. Organization Studies, 16(6), 1021-1046.
Bliese, P. D. (2000). Within-Group Agreement, Non-Independence, and
Reliability: Implications for Data Aggregation and Analysis. In K. J. Klein & S.
W. J. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in
Organizations: Foundations, Extensions, and New Directions (pp. 349-381). San
Francisco: Jossey-Bass.
Blumberg, M., & Pringle, C. D. (1982). The Missing Opportunity in
Organizational Research: Some Implications for a Theory of Work Performance.
Academy of Management Review, 7, 560-569.
Borgatti, S. P., & Cross, R. (2003). A Relational View of Information
Seeking and Learning in Social Networks. Management Science, 49(4), 432-445.
Brandon, D. P., & Hollingshead, A. B. (2004). Transactive Memory
Systems in Organizations: Matching Tasks, Expertise, and People. Organization
Science, 15(6), 633-644.
153
Burke, M. J., & Dunlap, W. P. (2002). Estimating Interrater Agreement
with the Average Deviation Index: A User's Guide. Organizational Research
Methods, 5(2), 159-172.
Campbell, J. P., & Pritchard, R. D. (1976). Motivation Theory in
Industrial and Organizational Psychology. In M. D. Dunnette (Ed.), Handbook of
Industrial and Organizational Psychology (pp. 63-130). Chicago: Rand McNally.
Cannon-Bowers, J. A., & Salas, E. (2001). Reflections on Shared
Cognition. Journal of Organizational Behavior, 22(2), 195-202.
Cannon-Bowers, J. A., Salas, E., & Converse, S. (1993). Shared Mental
Models in Expert Team Decision Making. Current issues in individual and group
decision making, 221-246.
Chan, D. (1998). Functional Relations among Constructs in the Same
Content Domain at Different Levels of Analysis: A Typology of Composition
Models. Journal of Applied Psychology, 83(2), 234-246.
Chen, G., & Klimoski, R. J. (2003). The Impact of Expectations on
Newcomer Performance in Teams as Mediated by Work Characteristics, Social
Exchanges, and Empowerment. Academy of Management Journal, 46(5), 591607.
Cooke, N. J., Kiekel, P. A., Salas, E., Stout, R., Bowers, C., & CannonBowers, J. (2003). Measuring Team Knowledge: A Window to the Cognitive
Underpinnings of Team Performance. Group Dynamics-Theory Research and
Practice, 7(3), 179-199.
Cooke, N. J., Salas, E., Cannon-Bowers, J. A., & Stout, R. J. (2000).
Measuring Team Knowledge. Human Factors, 42(1), 151-173.
Crossan, M. M., Lane, H. W., & White, R. E. (1999). An Organizational
Learning Framework: From Intuition to Institution. Academy of Management
Review, 24(3), 522-537.
Davenport, T. H., & Prusak, L. (1998). Working Knowledge: How
Organizations Manage What They Know: Harvard Business School Press.
De Dreu, C. K. W., & Weingart, L. R. (2003). Task Versus Relationship
Conflict, Team Performance, and Team Member Satisfaction: A Meta-Analysis.
Journal of Applied Psychology, 88(4), 741-749.
Denison, D. R., Hart, S. L., & Kahn, J. A. (1996). From Chimneys to
Cross-Functional Teams: Developing and Validating a Diagnostic Model.
Academy of Management Journal, 39(4), 1005-1023.
Dunlap, W. P., Burke, M. J., & Smith-Crowe, K. (2003). Accurate Tests
of Statistical Significance for R-Sub(Wg ) and Average Deviation Interrater
Agreement Indexes. Journal of Applied Psychology, 88, 356-362.
Dyer, J. H., & Nobeoka, K. (2000). Creating and Managing a HighPerformance Knowledge-Sharing Network: The Toyota Case. Strategic
Management Journal, 21(3), 345-367.
154
Eden, D. (1984). Self-Fulfilling Prophecy as a Management Tool:
Harnessing Pygmalion. The Academy of Management Review, 9(1), 64-73.
Edmondson, A. C. (1999). Psychological Safety and Learning Behavior in
Work Teams. Administrative Science Quarterly, 44(2), 350-353.
Edmondson, A. C., Bohmer, R. M., & Pisano, G. P. (2001). Disrupted
Routines: Team Learning and New Technology Implementation in Hospitals.
Administrative Science Quarterly, 46(4), 685-716.
Edmondson, A. C., Dillon, J. R., & Roloff, K. S. (2008). Three
Perspectives on Team Learning: Outcome Improvement, Task Mastery, and
Group Process. Academy of Management Annals, 1(1), 269-314.
Edwards, J. R. (1995). Alternatives to Difference Scores as Dependent
Variables in the Study of Congruence in Organizational Research. Organizational
Behavior and Human Decision Processes, 64(3), 307-324.
Edwards, J. R., & Parry, M. E. (1993). On the Use of Polynomial
Regression Equations as an Alternative to Difference Scores in Organizational
Research. Academy of Management Journal, 36(6), 1577-1613.
Ellis, A. P. J. (2006). System Breakdown: The Role of Mental Models and
Transactive Memory in the Relationship between Acute Stress and Team
Performance. Academy of Management Journal, 49(3), 576-589.
Entin, E. E., & Serfaty, D. (1999). Adaptive Team Coordination. Human
Factors, 41(2), 312-313.
Faraj, S., & Sproull, L. (2000). Coordinating Expertise in Software
Development Teams. Management Science, 46(12), 1554-1568.
Fiore, S. M., Salas, E., Cuevas, H. M., & Bowers, C. A. (2003).
Distributed Coordination Space: Toward a Theory of Distributed Team Process
and Performance. Theoretical Issues in Ergonomics Science, 4(3/4), 340-364.
Gersick, C. J. G., & Hackman, J. R. (1990). Habitual Routines in Task
Performing Groups. Organizational Behavior and Human Decision Process, 47,
65-97.
Gibson, C., & Vermeulen, F. (2003). A Healthy Divide: Subgroups as a
Stimulus for Team Learning Behavior. Administrative Science Quarterly, 48(2),
202-239.
Gilson, L. L., & Shalley, C. E. (2004). A Little Creativity Goes a Long
Way: An Examination of Teams' Engagement in Creative Processes. Journal of
Management, 30(4), 453-470.
Gladstein, D. L. (1984). Groups in Context: A Model of Task Group
Effectiveness. Administrative Science Quarterly, 29(4), 499-517.
Gladwell, M. (2002). The Talent Myth. The New Yorker, 28-33.
Gruenfeld, D. H., Mannix, E. A., Williams, K. Y., & Neale, M. A. (1996).
Group Composition and Decision Making: How Member Familiarity and
155
Information Distribution Affect Process and Performance. Organizational
Behavior and Human Decision Processes, 67(1), 1-15.
Gruenfeld, D. H., Martorana, P. V., & Fan, E. T. (2000). What Do Groups
Learn from Their Worldliest Members? Direct and Indirect Influence in Dynamic
Teams. Organizational Behavior and Human Decision Processes, 82(1), 45-59.
Guzzo, R. A., & Dickson, M. W. (1996). Teams in Organizations: Recent
Research on Performance and Effectiveness. Annual Review of Psychology, 47(1),
307-338.
Hackman, J. R. (1987). The Design of Work Teams. In J. Lorsch (Ed.),
Handbook of Organizational Behavior (pp. 315-342). Englewood Cliffs, NJ:
Prentice-Hall.
Hackman, J. R. (1988). Flight Crew Questionnaire. Cambridge, MA.
Hackman, J. R. (1990). Groups That Work (and Those That Don't):
Creating Conditions for Effective Teamwork. San Francisco: Jossey-Bass.
Haleblian, J., & Finkelstein, S. (1993). Top Management Team Size, Ceo
Dominance, and Firm Performance: The Moderating Roles of Environmental
Turbulence and Discretion. Academy of Management Journal, 36(4), 844-863.
Hinsz, V. B. (1990). Cognitive and Consensus Processes in Group
Recognition Memory Performance. Journal of Personality and Social
Psychology, 59(4), 705–718.
Hinsz, V. B., Tindale, R. S., & Vollrath, D. A. (1997). The Emerging
Conceptualization of Groups as Information Processors. Psychological Bulletin,
121(1), 43-64.
Hogg, M. A., & Terry, D. J. (2000). Social Identity and SelfCategorization Processes in Organizational Contexts. Academy of Management
Review, 25(1), 121-140.
Hollingshead, A. B. (1998). Retrieval Processes in Transactive Memory
Systems. Journal of Personality and Social Psychology, 74(3), 659-671.
Hollingshead, A. B. (2000). Perceptions of Expertise and Transactive
Memory in Work Relationships. Group Processes & Intergroup Relations, 3(3),
257-267.
Huber, G. P. (1991). Organizational Learning: The Contributing Processes
and the Literatures. Organization Science, 2(1), 88-115.
Ilgen, D. R., Hollenbeck, J. R., Johnson, M., & Jundt, D. (2005). Teams in
Organizations: From Input-Process-Output Models to IMOI Models. Annual
Review of Psychology, 56(1), 517-543.
James, L. R. (1982). Aggregation Bias in Estimates of Perceptual
Agreement. Journal of Applied Psychology, 67, 219-229.
156
James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating withinGroup Interrater Reliability with and without Response Bias. Journal of Applied
Psychology, 69, 85-98.
James, L. R., Demaree, R. G., & Wolf, G. (1993). R-Sub(Wg): An
Assessment of within-Group Interrater Agreement. Journal of Applied
Psychology, 78, 306-309.
Janis, I. L. (1972). Victims of Groupthink: A Psychological Study of
Foreign-Policy Decisions and Fiascoes. Boston: Houghton Mifflin.
Johns, G. (2006). The Essential Impact of Context on Organizational
Behavior. The Academy of Management Review, 31(2), 386-408.
Johnson-Laird, P. N. (1983). Mental Models: Toward a Cognitive Sceince
of Language, Inference, and Consciousness. Cambridge, MA: Harvard University
Press.
Jöreskog, K. G., & D. Sörbom (1996). Lisrel 8.0 User’s Reference guide.
Chicago: Scientific Software International.
Katz, R. (1982). The Effects of Group Longevity on Project
Communication and Performance. Administrative Science Quarterly, 27(1), 81104.
Katzenbach, J. R., & Smith, D. K. (2003). The Wisdom of Teams:
Creating the High-Performance Organization: Collins.
Keller, R. T. (1994). Technology-Information Processing Fit and the
Performance of R&D Project Groups: A Test of Contingency Theory. Academy of
Management Journal, 37(1), 167-179.
Kirkman, B. L., & Rosen, B. (1999). Beyond Self-Management:
Antecedents and Consequences of Team Empowerment. The Academy of
Management Journal, 42(1), 58-74.
Klimoski, R., & Mohammed, S. (1994). Team Mental Model: Construct or
Metaphor? Journal of Management, 20(2), 403-437.
Kogut, B., & Zander, U. (1992). Knowledge of the Firm, Combinative
Capabilities, and the Replication of Technology. Organization Science, 3(3), 383397.
Kozlowski, S. W. J., & Klein, K. J. (2000). A Multilevel Approach to
Theory and Research in Organizations: Contextual, Temporal and Emergent
Processes. In K. J. Klein & S. W. J. Kozlowski (Eds.), Multilevel Theory,
Research, and Methods in Organizations (pp. 3-90). San Francisco, CA: JosseyBass.
Lewis, K. (2003). Measuring Transactive Memory Systems in the Field:
Scale Development and Validation. Journal of Applied Psychology, 88(4), 587604.
157
Lewis, K. (2004). Knowledge and Performance in Knowledge-Worker
Teams: A Longitudinal Study of Transactive Memory Systems. Management
Science, 50(11), 1519-1533.
Liang, D. W., Moreland, R., & Argote, L. (1995). Group Versus
Individual Training and Group Performance: The Mediating Role of Transactive
Memory. Personality and Social Psychology Bulletin, 21(4), 384-393.
Libby, R., Trotman, K. T., & Zimmer, I. (1987). Member Variation,
Recognition of Expertise, and Group Performance. Journal of Applied
Psychology, 72(1), 81-87.
Liker, J. K. (2004). The Toyota Way: 14 Management Principles from the
World's Greatest Manufacturer: McGraw-Hill.
Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A Temporally
Based Framework and Taxonomy of Team Processes. Academy of Management
Review, 26(3), 356-376.
Marks, M. A., Zaccaro, S. J., & Mathieu, J. E. (2000). Performance
Implications of Leader Briefings and Team-Interaction Training for Team
Adaptation to Novel Environments. Journal of Applied Psychology, 85(6), 971986.
Mathieu, J. E., Goodwin, G. F., Heffner, T. S., Salas, E., & CannonBowers, J. A. (2000). The Influence of Shared Mental Models on Team Process
and Performance. Journal of Applied Psychology, 85(2), 273-283.
Mathieu, J. E., Heffner, T. S., Goodwin, G. F., Cannon-Bowers, J. A., &
Salas, E. (2005). Scaling the Quality of Teammates' Mental Models: Equifinality
and Normative Comparisons. Journal of Organizational Behavior, 26(1), 37-56.
Mohammed, S., & Dumville, B. C. (2001). Team Mental Models in a
Team Knowledge Framework: Expanding Theory and Measurement across
Disciplinary Boundaries. Journal of Organizational Behavior, 22(2), 89-106.
Mohammed, S., Klimoski, R., & Rentsch, J. R. (2000). The Measurement
of Team Mental Models: We Have No Shared Schema. Organizational Research
Methods, 3(2), 123-165.
Montoya-Weiss, M. M., Massey, A. P., & Song, M. (2001). Getting It
Together: Temporal Coordination and Conflict Management in Global Virtual
Teams. Academy of Management Journal, 44(6), 1251-1262.
Moreland, R. L., & Myaskovsky, L. (2000). Exploring the Performance
Benefits of Group Training: Transactive Memory or Improved Communication?
Organizational Behavior and Human Decision Processes, 82(1), 117-133.
Mowday, R. T., & Sutton, R. I. (1993). Organizational Behavior: Linking
Individuals and Groups to Organizational Contexts. Annual Review of
Psychology, 44(1), 195-229.
158
Murnighan, J. K., & Conlon, D. E. (1991). The Dynamics of Intense Work
Groups: A Study of British String Quartets. Administrative Science Quarterly,
36(2), 165-186.
Nahapiet, J., & Ghoshal, S. (1998). Social Capital, Intellectual Capital,
and the Organizational Advantage. Academy of Management. The Academy of
Management Review, 23(2), 242-266.
Nonaka, I. (1994). A Dynamic Theory of Organizational Knowledge
Creation. Organization Science, 5(1), 14-37.
Pelled, L. H. (1996). Demographic Diversity, Conflict, and Work Group
Outcomes: An Intervening Process Theory. Organization Science, 7(6), 615-631.
Peters, L. H., Fisher, C. D., & O'Connor, E. J. (1982). The Moderating
Effect of Situational Control of Performance Variance on the Relationship
between Individual Differences and Performance. Personnel Psychology, 35, 609621.
Peters, L. H., & O'Connor, E. J. (1980). Situational Constraints and Work
Outcomes: The Infouences of a Frequently Overlooked Construct. Academy of
Management Review, 5, 391-397.
Peterson, E., & Thompson, L. (1997). Negotiation Teamwork: The Impact
of Information Distribution and Accountability on Performance Depends on the
Relationship among Team Members. Organizational Behavior and Human
Decision Processes, 72(3), 364-383.
Rau, D. (2001). Knowing Who Knows What: The Effect of Transactive
Memory on the Relationship between Diversity of Expertise and Performance in
Top Management Teams. Unpublished Ph.D., University of Minnesota, United
States -- Minnesota.
Reagans, R., & Zuckerman, E. W. (2001). Networks, Diversity, and
Productivity: The Social Capital of Corporate R&D Teams. Organization Science,
12(4), 502-517.
Rentsch, J. R., & Hall, R. J. (1994). Members of Great Teams Think
Alike: A Model of Team Effectiveness and Schema Similarity among Team
Members. Advances in Interdisciplinary Studies of Work Teams, 1, 223-261.
Robertson, P. J., Roberts, D. R., & Porras, J. I. (1993). Dynamics of
Planned Organizational Change: Assessing Empirical Support for a Theoretical
Model. Academy of Management Journal, 36, 619-634.
Rosenthal, R. (1994). Interpersonal Expectancy Effects: A 30-Year
Perspective. Current Directions in Psychological Science, 3(6), 176-179.
Rouse, W. B., Cannon-Bowers, J. A., & Salas, E. (1992). The Role of
Mental Models in Team Performance in Complex Systems. IEEE Transactions on
systems, man and cybernetics, 22(6), 1296-1308.
159
Rouse, W. B., & Morris, N. M. (1986). On Looking into the Black-Box Prospects and Limits in the Search for Mental Models. Psychological Bulletin,
100(3), 349-363.
Rulke, D. L., & Rau, D. (2000). Investigating the Encoding Process of
Transactive Memory Development in Group Training. Group & Organization
Management, 25(4), 373-396.
Schmidt, F. L., & Hunter, J. E. (1989). Interrater Reliability Coefficients
Cannot Be Computed When Only One Stimulus Is Rated. Journal of applied
psychology, 74(2), 368-370.
Schneider, B. (1978). Person-Situation Selection: A Review of Some
Ability-Situation Interaction Research. Personnel Psychology, 31, 281-297.
Seger, C. A. (1994). Implicit Learning. Psychological Bulletin, 115(2),
163-196.
Senge, P. M. (1990). The Leader’s New Work: Building Learning
Organizations. Sloan Management Review, 32(1), 7-23.
Shrout, P. E., & Fleiss, J. L. (1979). Intraclass Correlations: Uses in
Assessing Rater Reliability. Psychol Bull, 86(2), 420-428.
Simon, H. A. (1997). Models of Bounded Rationality: MIT Press.
Smith-Jentsch, K. A., Mathieu, J. E., & Kraiger, K. (2005). Investigating
Linear and Interactive Effects of Shared Mental Models on Safety and Efficiency
in a Field Setting. Journal of Applied Psychology, 90(3), 523-535.
Stasser, G. L., & Stewart, D. D. (1992). Discovery of Hidden Profiles by
Decision-Making Groups: Solving a Problem Versus Making a Judgment. Journal
of Personality and Social Psychology, 63(3), 426-434.
Stasser, G. L., Taylor, L. A., & Hanna, C. (1989). Information Sampling
in Structured and Unstructured Discussions of Three- and Six-Person Groups.
Journal of Personality and Social Psychology, 57(1), 67-78.
Stasser, G. L., & Titus, W. (1987). Effects of Information Load and
Percentage of Shared Information on the Dissemination of Unshared Information
During Group Discussion. Journal of Personality and Social Psychology, 53(1),
81-93.
Stasser, G. L., Vaughan, S. I., & Stewart, D. D. (2000). Pooling Unshared
Information: The Benefits of Knowing How Access to Information Is Distributed
among Group Members. Organizational Behavior and Human Decision
Processes, 82(1), 102-116.
Steel, R. P., & Mento, A. J. (1986). Impact of Situational Constraints on
Subjective and Objective Criteria of Managerial Job Performance. Organizational
behavior and human decision processes, 37, 254-265.
Steiner, I. D. (1972). Group Process and Productivity: Academic Press
New York.
160
Stewart, D. D., & Stasser, G. L. (1995). Expert Role Assignment and
Information Sampling During Collective Recall and Decision Making. Journal of
Personality and Social Psychology, 69(4), 619-628.
Stewart, G. L. (2006). A Meta-Analytic Review of Relationships between
Team Design Features and Team Performance. Journal of Management, 32(1),
29-54.
Stewart, G. L., & Nandkeolyar, A. K. (2006). Adaptation and
Intraindividual Variation in Sales Outcomes: Exploring the Interactive Effects of
Personality and Environmental Opportunity. Personnel Psychology, 59, 307-332.
Stewart, G. L., & Nandkeolyar, A. K. (2007). Exploring How Constraints
Created by Other People Influence Intraindividual Variation in Objective
Performance Measures. Journal of Applied Psychology, 92(4), 1149-1158.
Sundstrom, E., de Meuse, K. P., & Futrell, D. (1990). Work Teams:
Applications and Effectiveness. American Psychologist, 45(2), 120-133.
Tajfel, H. (1982). Social Psychology of Intergroup Relations. Annual
Review of Psychology, 33(1), 1-39.
Tajfel, H., & Turner, J. C. (1986). The Social Identity Theory of
Intergroup Behavior. Psychology of Intergroup Relations, 2, 7-24.
Terborg, J. R. (1981). Interactional Psychology and Research on Human
Behavior in Organizations. Academy of Management Review, 6, 569-576.
Tesluk, P. E., & Mathieu, J. E. (1999). Overcoming Roadblocks to
Effectiveness: Incorporating Management of Performance Barriers into Models of
Work Group Effectiveness. Journal of Applied Psychology, 84, 200-217.
Thomas-Hunt, M. C., Ogden, T. Y., & Neale, M. A. (2003). Who's Really
Sharing? Effects of Social and Expert Status on Knowledge Exchange within
Groups. Management Science, 49(4), 464-477.
Thomas, J. B., & McDaniel Jr, R. R. (1990). Interpreting Strategic Issues:
Effects of Strategy and the Information-Processing Structure of Top Management
Teams. Academy of Management Journal, 33(2), 286-306.
Thompson, L., & Fine, G. A. (1999). Socially Shared Cognition, Affect,
and Behavior: A Review and Integration. Personality and Social Psychology
Review, 3(4), 278-302.
Tjosvold, D. (1985). Implications of Controversy Research for
Management. Journal of Management, 11(3), 21-37.
Van der Vegt, G. S., & Bunderson, J. S. (2005). Learning and
Performance in Multidisciplinary Teams: The Importance of Collective Team
Identification. Academy of Management Journal, 48(3), 532-547.
Van der Vegt, G. S., Emans, B. J. M., & Van de Vliert, E. (2001). Patterns
of Interdependence in Work Teams: A Two-Level Investigation of the Relations
with Job and Team Satisfaction. Personnel Psychology, 54(1), 51–69.
161
Watson, W., Michaelsen, L. K., & Sharp, W. (1991). Member
Competence, Group Interaction, and Group Decision Making: A Longitudinal
Study. Journal of applied psychology, 76(6), 803-809.
Webb, N. M. (1982). Student Interaction and Learning in Small Groups.
Review of Educational Research, 52(3), 421-445.
Wegner, D. M. (1987). Transactive Memory: A Contemporary Analysis of
the Group Mind. . In B. Mullen & G. R. Goethals (Eds.), Theories of Group
Behavior (pp. 185-208). New York: Springer-Verlag.
Wegner, D. M., Erber, R., & Raymond, P. (1991). Transactive Memory in
Close Relationships. Journal of Personality and Social Psychology, 61(6), 923929.
Weick, K. E. (1987). Organizational Culture as a Source of High
Reliability. California Management Review, 29(2), 112-127.
Weick, K. E. (1993). The Collapse of Sensemaking in Organizations: The
Mann Gulch Disaster. Administrative Science Quarterly, 38(4), 628-652.
Weick, K. E., & Roberts, K. H. (1993). Collective Mind in Organizations:
Heedful Interrelating on Flight Decks. Administrative Science Quarterly, 38(3),
357-381.
West, M. A. (2002). Sparkling Fountains or Stagnant Ponds: An
Integrative Model of Creativity and Innovation Implementation in Work Groups.
Applied Psychology: An International Review, 51(3), 355-387.
Wicks, A. C., Berman, S. L., & Jones, T. M. (1999). The Structure of
Optimal Trust: Moral and Strategic Implications. Academy of Management
Review, 24(1), 99-116.
Wiersema, M. F., & Bantel, K. A. (1992). Top Management Team
Demography and Corporate Strategic Change. Academy of Management Journal,
35(1), 91-121.
Wilson, J. M., Goodman, P. S., & Cronin, M. A. (2007). Group Learning.
Academy of Management Journal, 32(4), 1041-1059.
162
APPENDIX A: SURVEY
1. Please enter your team identification number
__________
(To be determined in consultation with the sponsoring organization, used for linking your
response with those of your teammates)
2. Please select your role
a. Team manager
b. Team manager
Note: If you selected team manager, please go to question number 14. If you are a team
member, please respond to the next question.
To be completed by the team members
Transactive Memory Systems (TMS)
3. For the following statements, please refer to the last completed project involving your
team. Please rate the following actions/behaviors in terms of your agreement with the
following statements on a scale of 1-5 (where 1 = strongly disagree 3 = neutral and 5=
strongly agree)
1
Strongly Disagree
2
Slightly
Disagree
3
Neutral
4
Slightly Agree
5
Strongly
Agree
(Specialization)
1. Each team member has specialized knowledge of some aspect of our project.
2. I have knowledge about an aspect of the project that no other team member has.
3. Different team members are responsible for expertise in different areas.
4. The specialized knowledge of several different team members was needed to
complete the project deliverables.
5. I know which team members have expertise in specific areas.
(Credibility)
6. In most cases, I was comfortable accepting procedural suggestions from other
team members.
7. In most cases, I trusted that other members’ knowledge about the project was
credible.
8. In most cases, I was confident relying on the information that other team members
brought to the discussion.
9. In most cases, when other members gave information, I wanted to double-check it
for myself. (reversed)
10. I did not have much faith in other members’ “expertise.” (reversed)
(Coordination)
11. Our team worked together in a well-coordinated fashion.
12. Our team had very few misunderstandings about what to do.
13. Our team needed to backtrack and start over a lot. (reversed)
14. We accomplished the task smoothly and efficiently.
163
15. There was much confusion about how we would accomplish the task. (reversed)
Team Mental Models (TMM)
Instructions: Below are some adjectives that could describe working in teams.
Operational Definitions:
1. Conflict Management: Establishing conditions to prevent team conflict and
working out interpersonal disagreements among team members.
2. Goal specification: Identification and prioritization of goals for task
accomplishment.
3. Monitoring progress toward goals: Tracking progress toward task
accomplishment.
4. Motivation and confidence building: Generating a sense of collective
confidence and motivation with regard to task accomplishment.
5. Strategy formulation: Developing alternative courses of action and choosing the
best course of action.
6. Team monitoring and backup behavior: Assisting team members in performing
their tasks by providing verbal feedback, helping or completing a task for them.
4. Think about the processes that were decided/established by your team/teammates
during the execution of the project. Now, based on your experience in the last completed
team project, please rank order the importance of the following team processes. Please
rank order them from 1-6 where 1= most important, 2 = second most important and so
on, 6 = least important.
_______
Conflict Management
_______
Goal specification
_______
Monitoring progress toward goal
_______
Motivation and confidence building
_______
Strategy formulation
_______
Team monitoring and backup behavior
164
Alternative TMM
5. For the following statements, please refer to the last completed project involving your
team. Please rate the following actions/behaviors in terms of your agreement with the
following statements on a scale of 1-5 (where 1 = strongly disagree 3 = neutral and 5=
strongly agree)
1
Strongly Disagree
2
Slightly
Disagree
3
Neutral
4
Slightly Agree
5
Strongly
Agree
1. We established rules to prevent, control, or guide team conflict before it occurred
and worked out interpersonal disagreements among team members.
2. We identified and prioritized goals for task accomplishment.
3. We tracked our progress toward task accomplishment.
4. We generated and worked on creating a sense of collective confidence and
engaged in motivating each other by building team cohesion.
5. We developed alternative courses of action for task accomplishment and chose
the best course of action.
6. We assisted team members to perform their tasks by providing verbal feedback or
coaching, helped others in carrying out actions, or completed a task for a
teammate.
6. For the following statements, please refer to the last completed project involving your
team. Please rate the following actions/behaviors in terms of your agreement with the
following statements on a scale of 1-5 (where 1 = strongly disagree 3 = neutral and 5=
strongly agree)
1
Strongly Disagree
2
Slightly
Disagree
3
Neutral
4
Slightly Agree
5
Strongly
Agree
Task Interdependence
1. I have to obtain information and advice from my colleagues in order to complete
my work.
2. I depend on my colleagues for the completion of my work.
3. I have a one person job; I rarely have to check or work with others.
4. I have to work closely with my colleagues to do my work properly.
5. In order to complete their work, my colleagues have to obtain information and
advice from me.
Reward for team performance
1. My performance review depends upon my performance as a member of the team.
2. My performance review depends upon the performance of the team.
3. Supporting my team is critical to advancement within the organization.
165
Organizational support
1. My team gets all the information it needs to do our work and plan our schedule.
2. It is easy for my team to obtain expert assistance when something comes up that
we don’t know how to handle.
3. My team is kept in the dark about current developments and future plans that may
affect its work.
4. My team lacks access to useful training on the job.
5. Excellent work pays off in this company.
7. For the following statements, please refer to the last completed project involving your
team. Please rate the following actions/behaviors in terms of your agreement with the
following statements on a scale of 1-5 (where 1 = strongly disagree 3 = neutral and 5=
strongly agree)
1
Strongly Disagree
2
Slightly
Disagree
3
Neutral
4
Slightly Agree
5
Strongly
Agree
Team viability
1. Members of my team care a lot about it, and work together to make it one of the
best.
2. Working with team members is an energizing and uplifting experience.
3. There is a lot of unpleasantness among members in the team. (reverse-coded)
4. Some members in the team do not carry their fair share of the overall workload.
(reverse-coded)
5. Sometimes, one of us refuses to help another team member out. (reverse-coded)
6. As a team, this work group shows signs of falling apart. (reverse-coded)
7. Every time we attempt to straighten out a member of the team, whose behavior is
not acceptable, things seem to get worse rather than better. (reverse-coded)
Team satisfaction
1. Generally speaking I am very satisfied with the team.
2. I frequently wish I could quit the team. (reverse coded)
3. I am generally satisfied with the work I do on the team.
Team Identity
8. I belong to:
a. only one work group (scored 3);
b. one primary work group, but also some secondary work groups (scored 2);
c. or more than one work group (scored 1). (Single team membership)
9. My primary work group:
166
a. consists mostly of members who are relatively permanent members of the
group (scored 3)
b. consists of some members who are relatively permanent and some
members who change frequently (scored 2)
c. consists of members who frequently change (scored 1). (Team member
permanence)
10. I would describe my primary work group as:
a. a group of members all working together as a single team (scored 3)
b. two or more subgroup of co-workers (scored 2)
c. or a collection of individual employees doing their own work (scored 1).
(Single-team functioning)
Team demographics
11. Now, please tell us a little bit about yourself.
1. Your age in years
2. No. of years with the present team
3. Number of years with the present organization
4. Number of members in your team
5. Your functional specialization (expertise)
12. Please select your gender
_________
_________
_________
_________
_________
_________ Male _________ Female
13. Please select the highest level of education you have completed.
1. Some high school (including high school graduate)
2. Some college coursework (including 2-year degrees)
3. 4-year college degree (e.g., BA/BS)
4. Some graduate coursework
5. Completed graduate degree (e.g., MA, MBA, MD, Ph.D.,
etc.)
14. Please select your industry.
1. Manufacturing
2. Banking/Finance/Accounting
3. Insurance / Real Estate / Legal
4. Federal (Including Military)/State / Local Government
5. Communications
6. Transportation / Utilities
7. Construction / Architecture / Engineering
8. Wholesale / Resale / Distribution
167
9. Education
10. Marketing / Advertising / Entertainment
11. Research / Development Lab
12. Computer/IT related
13. Business Consulting
THANK YOU FOR COMPLETING THE SURVEY!!
To be completed by the team managers
15. For the following statements, please refer to the last completed project involving your
team. Please rate the following actions/behaviors in terms of your agreement with the
following statements on a scale of 1-5 (where 1=somewhat below requirements 3=
neutral and 5 = consistently exceeds requirements).
1
Somewhat below
Requirements
2
Slightly below
Requirements
3
Neutral
4
Slightly above
Requirements
5
Consistently exceeds
Requirements
Team performance
1. This team meets or exceeds its goals.
2. This team completes its tasks on time.
3. This team makes sure that products and services meets or exceeds quality or
service standards.
4. This team responds quickly when problems come up.
5. This team successfully solves problems that slow down its work.
6. This team is a productive team.
Team learning
16. For the following statements, please refer to the last completed project involving your
team. Please rate the following actions/behaviors in terms of your agreement with the
following statements on a scale of 1-5 (where 1 = strongly disagree 3 = neutral and 5=
strongly agree)
1
Strongly Disagree
2
Slightly
Disagree
3
Neutral
4
Slightly Agree
5
Strongly
Agree
1. This team asks its internal customers (those who receive or use its work) for
feedback on its performance.
2. This team relies on outdated information or ideas. (reverse coded)
3. This team actually reviews its own progress and performance.
168
4. This team does its work without stopping to consider all the information team
members have. (reverse coded)
5. This team ignores feedback from others in the company when some tough
situation arises.
Team creativity
17. For the following statements, please refer to the last completed project involving your
team. Please rate the following actions/behaviors in terms of your agreement with the
following statements on a scale of 1-5 (where 1 = strongly disagree 3 = neutral and 5=
strongly agree)
1
Strongly Disagree
2
Slightly
Disagree
3
Neutral
4
Slightly Agree
5
Strongly
Agree
1. This team is methodical and consistent in the way it tackles problems. (reverse
coded)
2. This team is open to the implementation of new ideas and methods.
3. This team links ideas that originate from multiple sources.
4. This team is persistent in solving a problem even in novel situations.
5. This team searches for novel approaches not required at the time.
6. This team pays strict regard to the sequences and steps needed to complete a job.
(reverse coded)
Task type
18. This team is comprised of members who are mostly:
a) Homogenous
b) Heterogeneous
c) Neither of the above
19. The work done by this team is highly integral to the daily operations of the
organization.
a) Yes
b) No
c) Can’t say
20. This team can best be classified as:
a) A homogenous team engaged in production or service activities that are highly
integral to the daily operation of the organization e.g. Sales team, assembly team.
b) A homogenous team engaged in decision making activities that are loosely
connected to the daily operation of the organization e.g. Employee involvement
groups, Quality control circles.
c) A heterogeneous team of specialized individuals working in an unpredictable
environment and engaged in activities highly integral to the daily operation of the
organization e.g. surgery team, flight crew.
169
d) A heterogeneous team comprised of specialized professionals assigned to a
project and not directly involved in the daily operation of the organization e.g.
R&D, New product development team.
e) Other (please specify).
If you selected other, please specify: _________________________________
21. Please describe the nature of task last completed by your team in a sentence or two.
________________________________________________________________________
________________________________________________________________________
THANK YOU FOR COMPLETING THE SURVEY!!
170
APPENDIX B: INFORMED CONSENT DOCUMENT
Project Title:
How do teams learn? Shared Mental Models and Transactive
Memory Systems as determinants of team effectiveness
Research Team: Amit Nandkeolyar, MBA
This consent form describes the research study to help you decide if you want to
participate. This form provides important information about what you will be asked to do
during the study, about the risks and benefits of the study, and about your rights as a
research subject.
• If you have any questions about or do not understand something in this form, you
should ask the research team for more information.
• You should discuss your participation with anyone you choose such as family or
friends.
• Do not agree to participate in this study unless the research team has answered
your questions and you decide that you want to be part of this study.
WHAT IS THE PURPOSE OF THIS STUDY?
This is a research study towards the completion of a PhD dissertation. I am inviting you
to participate in this research study because you are employed in a commercial
organization and routinely work in teams.
The purpose of this research study is to gain an understanding of team effectiveness in
real world settings. Further, I seek to examine if team learning explains a team’s ability to
perform successfully.
HOW MANY PEOPLE WILL PARTICIPATE?
Approximately 250-300 people will take part in this study conducted by researchers at
the University of Iowa.
HOW LONG WILL I BE IN THIS STUDY?
If you agree to take part in this study, your involvement will last for 15 minutes.
WHAT WILL HAPPEN DURING THIS STUDY?
You will be asked to complete a survey either online or in a paper-pencil format. You
will be asked to enter the team identification number which will be provided to you by
the researcher in the e-mail inviting you to participate in the survey. You will then select
your role as a team member or a team manager.
The survey will ask you to answer certain questions about your experience in working
with teams. Team members will also be asked to provide our age, sex, number of years
171
with the team and organization, number of members in your team, highest educational
degree obtained, and your specialization/expertise. You are free to skip any questions that
you would prefer not to answer.
WHAT ARE THE RISKS OF THIS STUDY?
There are no foreseeable risks to participating in this study.
WHAT ARE THE BENEFITS OF THIS STUDY?
We don’t know if you will benefit from being in this study. However, we hope that, in the
future, other people might benefit from this study because we will learn about team
processes in commercial settings.
WILL IT COST ME ANYTHING TO BE IN THIS STUDY?
You will not have any costs for being in this research study.
WILL I BE PAID FOR PARTICIPATING?
You will not be paid for being in this research study.
WHO IS FUNDING THIS STUDY?
The University and the research team are receiving no payments from other agencies,
organizations, or companies to conduct this research study.
WHAT ABOUT CONFIDENTIALITY?
We will keep your participation in this research study confidential to the extent permitted
by law. However, it is possible that other people such as those indicated below may
become aware of your participation in this study and may inspect and copy records
pertaining to this research. Some of these records could contain information that
personally identifies you.
• federal government regulatory agencies,
• auditing departments of the University of Iowa, and
• the University of Iowa Institutional Review Board (a committee that reviews and
approves research studies)
To help protect your confidentiality, we will not put your name on any part of the
questionnaire. Your responses will be available only to members of research team. The
hard copy, if any, of data will be stored in locked file cabinets in locked offices. Any
research information stored on computers will be stored in password protected computer
files. The team ID Code assigned to your team will be linked to identifying information
about your organization and team. The list linking the team identification code and your
172
organization/team will be stored in a separate location that is accessible only to the
researchers.
The overall results of the study will be provided to your organization. If we write a
report or article about this study or share the study data set with others, we will do so in
such a way that you cannot be directly identified.
IS BEING IN THIS STUDY VOLUNTARY?
Taking part in this research study is completely voluntary. You may choose not to take
part at all. If you decide to be in this study, you may stop participating at any time. If
you decide not to be in this study, or if you stop participating at any time, you won’t be
penalized or lose any benefits for which you otherwise qualify.
WHAT IF I HAVE QUESTIONS?
We encourage you to ask questions. If you have any questions about the research study
itself, please contact: Amit Nandkeolyar, 319-335-1504 (O), 319-354-8161 (H).
Alternatively, you may contact Dr. Greg Stewart at 319-335-1947 (O) who is the chair of
the dissertation committee. If you experience a research-related injury, please contact:
Amit Nandkeolyar at 319-335-1504 (O), 319-354-8161 (H).
If you have questions, concerns, or complaints about your rights as a research subject or
about research related injury, please contact the Human Subjects Office, 340 College of
Medicine Administration Building, The University of Iowa, Iowa City, Iowa, 52242,
(319) 335-6564, or e-mail [email protected]. General information about being a research
subject can be found by clicking “Info for Public” on the Human Subjects Office web
site, http://research.uiowa.edu/hso. To offer input about your experiences as a research
subject or to speak to someone other than the research staff, call the Human Subjects
Office at the number above.
By completing and submitting the surveys, you are consenting to the use of your
responses in our research study. To begin the survey, please GO TO THE NEXT PAGE.
You may stop your participation at any time by closing the browser window.
173
APPENDIX C: GLOSSARY
Information processing: Information processing in teams has been defined as the degree
to which information, ideas, or cognitive processes are shared, and are shared among
group members and how this sharing of information affects both individual and group
level outcomes.
Shared mental models (SMM): Organized understanding of relevant knowledge that is
shared amongst team members
Task (Shared) Mental Model: SMM typology focusing on the knowledge about
equipment and tools & task procedure, strategies and environmental cues impacting task.
Team (Shared) Mental Model (TMM): SMM typology focusing on the knowledge about
teammate characteristics, team roles and team interaction patterns.
Team learning: Process by which team members seek to acquire, share, refine, or
combine task relevant knowledge through interaction with one another.
Transactive Memory Systems (TMS): A combination of an individual’s knowledge and a
shared awareness of who knows what.
TMS Coordination: Team members’ ability to work together efficiently.
TMS Credibility: Team members’ beliefs about the accuracy of each member’s
knowledge and provides evidence that team members trust each other.
TMS Specialization: Team members’ level of knowledge differentiation